Foundry Practice Foseco

THE AUTHORITATIVE MAGAZINE FOR FOUNDRY ENGINEERS

SYNTHETIC SANDS

AUTOMATIC POURING

STEEL FILTRATION

JANUARY 2013

C O A T I N G S F I LT R A T I O N F E E D I N G S Y S T E M S M E LT S H O P R E F R A C T O R I E S M E T A L T R E A T M E N T B I N D E R S C R U C I B L E S

FOUNDRY PRACTICE 258

The use of synthetic sand as a moulding media within foundries is increasing due to the rising cost of specialist sands such as zircon and chromite and the benefits of utilising a single sand to increase reclamation rates and reduce the costs associated with dumping used sand. The article provides an overview of the performance of different grades of synthetic sand when compared with silica, zircon and chromite in terms of properties and casting performance. Additionally tests are carried out to determine whether the application performance of refractory coatings are adversely affected by the change of moulding media.

THE AUTHORITATIVE MAGAZINE FOR FOUNDRY ENGINEERS

Automatic mould pouring systems have gained from successive innovations and efforts to thoroughly understand the process of pouring with a stopper, so that today they are very close to fully meeting founders’ expectations.

This paper explains why some foundries remain sceptical about automatic pouring, before focusing on the innovative approach proposed by SERT Metal in the field of advanced pouring control, with a predictive controller that makes it possible to generate significant savings.

01

AUTOMATIC POURING Predictive control: one step ahead in automatic pouring with stopperAuthor/s: C. Debray, D. Alain, F. Montegu and X. Rabec. SERT Metal

STEEL FILTRATION The use of fi ltration to improve the quality of large steel railroad castingsAuthor/s: E.O. Chertovskikh, Altai State Technical University, N. V. Svalov & V. A. Polenov, Foseco LLC, S. V. Kushakov, OJSC Altaivagon

This paper demonstrates how the use of carbon-bonded, ceramic foam filters in the gating systems of railcar bogie side frame castings brings about an improvement in casting quality, due to a reduction in melt flow turbulence during mould filling. Use of the filters also protects the metal from secondary oxidation during pouring and inhibits slag caused by gating system erosion. The experiments conducted show how an increase in steel purity is obtained through the removal of non-metallic inclusions, namely oxides, sulphides and alumino-silicates.

08

05

02 SYNTHETIC SANDSCharacteristics of synthetic sand and application of refractory coatingsAuthor/s: T. Okada and T. Ikeuchi, Foseco Japan Ltd.

JANUARY 2013

FOUNDRY PRACTICE 258

CMYKGrey : 0 / 0 / 0 /85Red : 0 / 100 / 96 / 0

02

IntroductionIn recent years there has been a drive to increase the re-use of foundry sands to limit the cost of new sand purchases and to minimise both the cost of dumping used sand and to limit environmental impact. Traditional foundry sands systems are often based on silica sand, with speciality sands, such as chromite, being used in areas of the mould where higher refractoriness or thermal stability is required. However these additions of speciality sands (specifically chromite) can have a detrimental effect on the performance of the reclaimed sand reducing its refractoriness and increasing the potential of sand fusion and metal penetration, and should be effectively separated before re-use of both the reclaimed silica and chromite components. [1]

Additionally the physical characteristics of silica sand such as grain shape, angularity and porosity can limit the effectiveness of mechanical reclamation processes by preventing the easy removal of the binder without damaging or fracturing the sand grain. High intensity scrubbing of the sand to effectively remove residual binder can lead to a reduction in the average grain size of the reclaimed sand. This will then require increased binder additions to subsequently re-bond the processed sand and increase dust levels that can, if not controlled, contribute to airborne particulates and a respirable silica dust hazard. [2]

The use of synthetic sands offer the opportunity to eliminate the use of both silica and special refractory sands, and provide good yield during the reclamation process:

• Superior refractory properties compared with silica sand

• Low thermal expansion

• High strength, resistant to breakage during reclamation

• Spherical shape, enabling easier removal of residual binder

Synthetic sands have a different composition and form compared with both silica and chromite sand, and it follows that the performance of the mould coating may also differ in terms of application properties and subsequent casting surface finish.

This article considers the characteristics of synthetic sands when compared with both silica, zircon and chromite sand and assesses the application performance of the mould coating.

The fundamental characteristics of synthetic sandsThe characteristics of commercially used silica, zircon, chromite and synthetic sands are shown in table 1. On determining the composition of the synthetic sands by means of X-ray diffraction they were all found to be alumina silicates. Furthermore, as shown in figure 1, all of the synthetic sands have a distinct spherical form.

Characteristics of synthetic sand and application of refractory coatings

Sand Silica Sand

Zircon Sand

Chromite Sand

Synthetic Sand 1

Synthetic Sand 2

Synthetic Sand 3

Density(g/cm3)

1.7 3.0 2.9 1.8 2.1 1.7

AFS 56 116 53 70 75 68

Composition SiO2 ZrSiO4 FeCr2o4

MgCr2O4

Al6Si2O13 Al6Si2O13 Al6Si2O13

Components SiO2 99% ZrO2 66%

SiO2 33%

Cr2O3 46%

FeO 26%

MgO 9%

Al2O3 56%

SiO2 39%

Fe2O3 2%

Al2O3 63%

SiO2 27%

Fe2O3 3%

Al2O3 57%

SiO2 34%

Fe2O3 2%

Table 1. Properties of various types of sand

Figure 1. Form of the sand

Characteristics of synthetic sand and application of refractory coatings

Issue 25803

The heat conducting properties of synthetic sandsTo measure the heat conducting properties of the different sands, test pieces were placed on a heat source as shown in figure 2, and temperature at a set point within the test piece was recorded over time using a thermocouple and data recorder. The measured results at a number of fixed time points are shown in table 2.

When comparing the recorded temperature within the test pieces after they had been located on the heat source for 5 minutes, the temperature was lower with all of the synthetic sands than with silica sand and that the highest temperature was achieved with the zircon sand. From these results it was concluded that the thermal conductivities of the synthetic sands were lower than that of silica sand and significantly lower than that of zircon sand.

Refractory performance of synthetic sandsCasting tests were carried out with uncoated moulds in order to evaluate the refractory performance of the synthetic sands. A furan mould was constructed to produce a hexagonal cross-sectioned casting with each side measuring 350mm wide by 720mm high, and incorporating two cylindrical cores in each side, see figure 3. The cores that formed each side were made of different base sands including silica, zircon, chromite and synthetic sands. The casting was poured using standard carbon steel at 1650°C, and the resultant casting weighed 110kg.

The casting results are shown in figure 4. The casting surface adjacent to the silica sand exhibits burn-on, both on the hexagonal face and within the cylindrical indentation, whereas adjacent to the zircon sand the casting surface was smooth. For the synthetic sands it was observed that both samples 1 & 2 showed some burn-on in the cylindrical indentation and on the hexagonal face. The third sample showed a clean defect free surface. The chromite sand had a similar performance to the synthetic sand samples 1 & 2, with a rougher surface finish than either zircon or synthetic sand 3 and some minor burn-on defects.

Measurement Conditions

Test Piece φ50 x 50

Furan Resin 1.5%

Curing Agent 0.75%

Heat Source Temperature 1550OC

Position at which the temperature was measured

Within the test piece (10mm from the heat source)

Figure 2. Outline drawing of the set-up for temperature measurement

Temperature (OC)

Holding Time 1 Minute 3 Minutes 5 Minutes

Silica Sand 40 155 280

Zircon Sand 50 180 315

Chromite Sand 20 150 270

Synthetic Sand 1 25 150 260

Synthetic Sand 2 20 95 200

Synthetic Sand 3 20 135 250

Table 2. Temperature at fi xed point within test piece versus time

Figure 3. Hexagonal casting test piece

Figure 4. Casting results

04

Consequently when a switchover from conventional sand is made it is recommended that a coating should be selected that can compensate for the different properties of synthetic sands when compared with conventional sands. By establishing the correct synthetic sand, binder and coating combination it is possible to utilise a single sand type, eliminating the use of expensive speciality sands for mould facing. This will allow for an optimised sand reclamation process that minimises new sand purchases and dumping costs, whilst additionally eliminating the formation of silica dust that may contribute to a respirable hazard.

References[1] Influence de la presence de chromite – Poyet P and Chevriot R – Fondrie 1980 Vol.35 p93-103

[2] OSHA Fact Sheet – Crystalline Silica Exposure (Health Hazard Information)

Application of mould coatings to synthetic sand mouldsTo establish whether the application performance of a mould coating could be adversely affected by the mould substrate, the differences in applied layer thickness and the penetration of the coating into the mould surface was determined for each substrate. A standard zircon-based coating with an ethanol solvent was used (ISOMOL* 310PE).

The coating was adjusted to 60 Baume and applied to a standard test-piece by dipping, and dried by a subsequent ignition of the solvent. The results are shown table 3.

The coating penetration depth into the mould surface was significantly lower for the synthetic sands when compared with silica sand, whilst layer thickness was equivalent in all these cases. The zircon sand had a significantly lower penetration depth, which also resulted in an increased average layer thickness. Comparing the synthetic sands, there was a marked variation in the penetration depths with synthetic sand 3 being significantly lower than the other two samples; this can be attributed to the slightly lower AFS number, the particle size distribution, form and composition.

ConclusionsThe investigation of widely used commercial synthetic sands has shown that the bonded sand has a lower thermal conductivity than silica, zircon or chromite sand providing a more insulating moulding material. The refractoriness of the synthetic sand under casting conditions is significantly better than that of silica sand, but the indications are that it does not perform as well as zircon sand and the performance can be dependent on the particular grade of material. Standard coatings can be applied to synthetic sand moulds without problems and similar layer thicknesses are achieved without modification to the coating dilution. However, penetration into the mould surface can be less than with conventional sand types, and is effected further by the specific grade of synthetic sand.

According to these results it is possible that when using synthetic sands that the solidification of the cast material will be retarded and that hot spots may be formed in areas not previously observed with conventional sand types. Moreover, since the penetration of the coating into the mould surface is reduced it is possible that defects such as burn-on may arise in parts where they have not occurred previously.

Sand Silica Sand

Zircon Sand

Chromite Sand

Synthetic Sand 1

Synthetic Sand 2

Synthetic Sand 3

AFS 56 115 53 70 75 68

Dry Layer (µm)

150-200 200 100-200 150-200 150-200 150-200

Penetration Depth (µm)

3500 500 2500 2000 2500 1000

Table 3. Comparison of applied coating layer to different sand substrates

Characteristics of synthetic sand and application of refractory coatings

CMYKGrey : 0 / 0 / 0 /85Red : 0 / 100 / 96 / 0

05 Issue 258

SERT MetalIn December 2011, Vesuvius acquired SERT Metal.

Located in Lyon, France, SERT Metal is a worldwide leader in the development, manufacture and marketing of high performance systems for the automation of casting processes of molten metal.

Launched in 1965, SERT Metal has developed industrial systems and equipment for the improvement of metal flow control. SERT Metal solutions are applied to existing or new continuous steel casters or pouring machines mostly, but not only, using stopper flow control.

SERT Metal’s skills cover every main type of process related to the pouring or casting of liquid metal. It supplies complete automated turnkey solutions, specific products and assistance for its customers to optimise productivity, quality and operational safety.

SERT Metal has an in-depth extensive knowledge in various caster types, argon flow control, temperature management and inoculation.

IntroductionAutomatic mould pouring systems have benefited from successive innovations and efforts to thoroughly understand the process of pouring molten iron using stopper technology; today we are very close to fully meeting foundry users expectations. These systems can now be applied in the vast majority of situations and their performance is constantly improving, providing reduced and stabilised pouring times, optimised dosing and adherence to process control requirements.

There are various suppliers in this market, who propose solutions which are similar in terms of equipment. Electric actuators for stopper driving and optical sensors are well established. The difference between the various systems available is associated with intelligence, autonomy and response to the variety of situations found in the metal pouring process (see figure 1).

This paper will start with a definition of “automatic pouring”, as the phrase is often misunderstood, and even sometimes misused. We will also try to explain why some foundry users

remain sceptical about it. We will then focus our attention on the innovative approach proposed by SERT Metal in the field of advanced pouring control, with a novel predictive controller that makes it possible to generate significant savings. The example presented later will illustrate the relevance of this approach and the value it can deliver.

Automatic pouring: defi nitionAn automatic system should function autonomously, and should do so for a long period of time. Operator’s assistance should only be required to input the setpoint parameters. The system should then manage the pouring operation according to the process requirements previously input into the system. These process requirements will have been previously defined by the experts of the process. Analogy can be made with an airline pilot who inputs the destination into his automatic pilot; the automatic pilot then manages the flying of the plane with only occasional monitoring from the pilot. A similar situation should be experienced in the foundry in association with automatic pouring.

Automated pouring without sensors

Due to the complexity of supplying a fully automatic pouring solution, some pouring machine manufacturers offer pouring systems which they call automatic, but what they actually supply are only “Teach In systems”, i.e. systems that can endlessly repeat a stopper opening curve, without any control. Logically, they should be referred to as semi-automatic pouring, which is the proper term for this technology, the limitations of which are well known.

Predictive control: one step ahead in automatic pouring with a stopperT.K. Waupaca Plant (WI, USA) case study and results

Figure 1. Mould pouring process and fl uctuations

06

In addition, several options are available that contribute to improve iron control, traceability and quality:

• automatic positioning of the pouring machine

• measurement and control of the metal level in the channel

• in-stream temperature measurement

• in-stream inoculant feeding

• in-stream inoculation checking

• adjustment of iron temperature at the pouring point.

Predictive control approachUntil a few years ago the UCERAM system functioned through the evolution of some key attributes to pouring, based on “self teaching” mould after mould. The principle is to react to variation in the height of the metal in the pouring cup, and adjust the stopper actuation to maintain a constant metal height in the pouring cup. For a mould that had a long pouring time the stopper would be moving to correct the metal height. This system is therefore designed to compensate for variations in the flow requirement in a reactive manner. In addition, there is a significant lapse in time between the stopper movement and the resultant adjustment of metal height in the pouring cup: this time lapse can be up to 1 second and can generate a loss of control and “pumping” in the pouring cup.

It was considered that a system that can anticipate the variations and compensate for them before they even occur would be highly beneficial. An example of where this kind of system could be really valuable is in the pouring of nodular iron castings which usually have a high feeder neck. Consequently SERT Metal undertook some research work in 2007 in cooperation with an advanced Automation Research Centre. The initial purpose was to reduce the number of settings that had to be input by an operator on the pouring machine. The work naturally led to the design and development of a controller that can detect the ideal pouring requirements (pouring schedule) throughout the casting of a given mould and automatically input it into itself as the setpoint to follow mould after mould.

The ideal pouring schedule is defined by the UCERAM system by initially pouring a few moulds using a classic controller and saving the pouring schedule in the system’s memory. This schedule is then used by the UCERAM system and constantly improved as moulds are poured, by corrections in real-time, until the ideal pouring conditions and schedule for the mould have been defined to maintain a consistent metal height in the pouring cup. This system can now be defined as predictive and compensates for any mould related variations in metal height in the pouring cup before they occur.

The modern UCERAM system now integrates a predictive and reactive approach to providing a novel solution to ideal mould filling. The predictive approach provides the ability to define and deliver the ideal molten metal pouring condition for a specific pattern; the reactive approach allows the system to compensate for fluctuations in the pouring process caused by metal temperature, wear or clogging of the stopper and nozzle, sand humidity etc. The UCERAM system can now provide a virtually perfect pouring schedule to each casting with very little operator input.

The most important of these limitations relates to operator competence; the pouring quality that a “Teach In system” will deliver is dependent on the quality of the pouring operator and their input. A “Teach In system” can temporarily free the operator, provided it is working with a high quality stopper actuator and on a steady process, but there are many variables inherent to the process that will need operator management. These variables include clogging of the nozzle/stopper valve, fluctuations in metalostatic height and variations in iron temperature that will cause the operator to have to return to the control station.

Semi automatic pouring with sensors and basic control

Equipment is available, described as “automatic pouring systems” where the control is undertaken by a sensor. For many of these, the adjustments necessary to obtain quality pouring are too numerous and complicated to be done properly, or it becomes a “Teach In” type system which is later controlled by the sensor.

In both those cases, human involvement remains very important in defining the quality of operation of the machine, and should not be defined as an automatic system.

The confused language and play with words which we have mentioned above partly explains why some foundry users are still sceptical and dubious about automatic pouring.

Full automatic pouring with sensors and predictive control

The measurement and control of molten metal flow in iron & steel plants and foundry processes, has been SERT Metal’s speciality for about 30 years. Among our solutions dedicated to foundries, we offer a system for automatic mould pouring that meets the criteria defined above, i.e. autonomous and loaded with complex algorithms that are fit to respond to process variations without requiring constant attention from the operator. The requirement on the operator is limited to inputting some pouring targets specific to the pattern under production, particularly the metal level in the cup after pouring. The system then self-adapts and quickly adjusts to meet those objectives together with maintaining standard quality requirements, such as rapid opening and closing, minimised and steady pouring time, and optimised dosing of additions.

Since its launch in 1996, about 80 of SERT Metal’s UCERAM* automatic mould pouring units have been sold to approximately 40 foundries in Europe, Asia and the USA. UCERAM comprises of the following elements:

• an electric actuator for stopper driving

• an advanced pouring controller

• an optical sensor, based on a multi measurement image analysis device that can provide the controller with real-time information on the pouring area (iron level in the cup, stream width, nozzle leakage etc).

Predictive control: one step ahead in automatic pouring w

ith a stopper.

07 Issue 258

Case StudyWaupaca Plant 4 is a foundry located in Marinette (Wisconsin, USA). It belongs to ThyssenKrupp Waupaca group, which has 6 foundry plants in the US, producing automotive parts.

Waupaca Plant 4 produces nodular iron safety parts on 6 vertical moulding lines, each equipped with a pressure furnace. Between 2003 and 2005, Waupaca Plant 4 equipped their 6 lines with one automatic pouring system UCERAM each (on the whole, Waupaca group has 20 UCERAM systems on 5 plants).

Early 2011, Waupaca Plant 4 decided to upgrade the UCERAM system. The latest generation, including the predictive control technique, has been implemented on the six moulding lines.

After a few months, production results demonstrate a very positive impact, at several levels:

• operators’ involvement has been drastically reduced and simplified. Considering that the same operators are busy with coring and driving the moulding machine and the furnace, that gain is all the more significant in that plant

• pouring times have been reduced and stabilised, and the end levels are clearly closer to the setpoint, as shown in figure 2. In this case, the standard deviation on pouring time (in red) goes down from 0.8s to 0.3s and the standard deviation on iron level in the cup (in blue) goes down from 15 mm (37%) to 5 mm (14%)

• the new system as a whole is very simple to troubleshoot. 100% of the calls from operation to maintenance are now related to a refractory issue or improper mounting of the stopper/nozzle part

• system reliability is at top level: since the upgrade, there has been no downtime due to pouring machine maintenance.

Tim ALLEN, Plant Manager at TK Waupaca Plant 4, testifies:

“All of our pouring related quality and productivity indicators show signifi cant and steady improvement since the upgrade.

As a foundry, we can see the impact of this upgrade on our bottom line.

This particular project has not only met but also exceeded our expectations in a way that few projects do.”

ConclusionThis application example shows that this predictive approach applied to foundry automatic pouring brings very significant gains. Since the first trials with this innovative technique, some ten installations have been commissioned and have obtained a very quick return on investment (usually less than 6 months).

The predictive controller is particularly adapted when the pouring time determines the moulding rate or when the pattern being produced is subject to important flow variations in the course of pouring.

When it is implemented, the immediate result is the repetitive production of quality parts, while minimising the pouring times and the quantity of metal poured.

This new approach is going one step ahead in pouring control. The autonomy it brings to the system provides real meaning to the phrase “automatic pouring”.

This paper was originally published in Cast Metal & Diecasting Times, October/November 2011.

Figure 2. Comparison with/without predictive control on a sample of 1800 moulds of the same pattern

CMYKGrey : 0 / 0 / 0 /85Red : 0 / 100 / 96 / 0

08

BackgroundThe Rubtsovsk branch of the company OJSC Altaivagon produces castings for railway truck bogies. One of the more difficult components to manufacture is the bogie side frame, due to the high standards required. When in use, bogie side frames are subject not only to static and dynamic loads, but also to torsional effects associated with the turning motion of the trucks. Dynamic loading is primarily cyclical in nature.

Owing to a serious shortage of side frame components on international markets, technology is now being called for to support the sustainable production of castings capable of complying with strict inspection specifications.

Ceramic foam filters first appeared in the 1970s and have primarily been used for the filtration of cast iron and non-ferrous alloys. Initially used outside Russia, the filters were first introduced to Russia at the beginning of the year 2000. The development of filtration in the casting of steels has not, however, been as simple as that of cast iron and the non-ferrous alloys. Many companies have tested filters, but many production facilities have decided against their use, due to poor results. The main problems encountered have centred on short castings and filter breakdown. This is due to the fact that steels, by contrast to cast iron and non-ferrous alloys, are strongly contaminated by the products of oxidation and possess low slag fluidity. These difficulties have limited adoption of filters over the last ten years. During this period, there have been developments in the adaptive gating systems for the filters and issues relating to steel refinement have been solved.

In relation to the purity of the metal castings, the first step is to establish the source of any incoming non-metallic inclusions. These can be divided into two main groups – endogenous inclusions and exogenous inclusions. Exogenous inclusions are foreign bodies, typically sand particles, mould and lining materials and furnace and ladle slag. Indigenous inclusions are the products of chemical oxidation reactions, which take place during the production and pouring of the metal. Indigenous inclusions are represented by silicates, oxides, nitrides, sulphides and their compounds [1].

The reducing agent primarily used in steel casting is aluminium, which has a strong affinity for oxygen and reduces iron from its oxide to give aluminium oxide. In the iron oxide solubility diagram (Figure 1), it can be seen that

the lower the carbon content of a steel, the higher its iron oxide solubility. So, reactions involving 0.8% carbon steel require 23 grams/tonne of aluminium and produce 45 grams of aluminium oxide, whereas those involving 0.16% carbon steel require a minimum of 200 grams of aluminium and produce 376 grams of aluminium oxide. It follows that the lower the carbon content of the steel, the greater the amount of reducing agent required and the greater the amount of oxidation products formed [2].

It should be noted, that the process of secondary oxidation and consequent reduction giving rise to the formation of inclusions takes place in the mould cavity during the pouring of the melt. The intensity of this process is dependent upon the metal flow in the mould cavity. The flow is dictated by the gating system configuration and the point of delivery of the metal into the casting.

Research undertaken in the USA at the end of the 1980s at 14 casting facilities specialising in the production of carbon and low-alloy steels [1] focused on the testing of 500 specimens cut from castings. In 83% of the specimens tested, the non-metallic macro-inclusions found in the castings were the products of secondary oxidation, i.e. oxidation of the metal at the time of its pouring into the mould.

The use of fi ltration to improve the quality of large steel railroad castings

The use of fi ltration to improve the quality of large steel railroad castings

Figure 1. Diagram showing oxygen solubility as a function of carbon content (1620 C)

09 Issue 258

The inclusions detected had magnitudes of up to 10mm. The results suggest that in order to reduce the incidence of non-metallic inclusions, it is first of all necessary to prevent or minimise the process of secondary oxidation of the metal in the mould. Provided the associated gating system is properly constructed, a filter is able to stop the passage of non-metallic inclusions coming from the ladle and inhibit the formation of oxide inclusions in the mould.

The application and benefi t of fi ltersFoseco STELEX* PrO (Figure 2) is a range of carbon-bonded ceramic-foam filters. These filters have the following special physical properties: low heat storage capacity, high durability in the heated state, high flame resistance, low thermal expansion and tolerance to thermal shock. These special properties make the filters suitable for use in the casting of low-carbon steels.

In order to use the filter in gating systems, Foseco’s specialists have developed an optimum gating system, which ensures that filter loading is optimised during the pouring of the melt. The gating system also minimises flow turbulence and, consequently, reduces oxidation of the steel in the mould cavity – a fact confirmed by the findings of the MAGMA* programme.

The results of the experimental work undertaken demonstrate how the chemical make-up of the casting is altered. A selection of tests was carried out on 3 melts poured from bottom-pour ladles to determine their chemical content. The chemical make-up of the castings was determined on templates cut from a journal box hatch. Comparison of the results obtained from the chemical analysis suggests that no carbonisation effects are present when using a filter (Table 1). The tests undertaken demonstrate a slight reduction in the carbon content of the casting of 0.01 – 0.02%.

As a result of the molten metal flowing through the filter, a 3- to 9-fold reduction in the Ca content of the casting is observed, when compared with the results of the chemical analysis of the melt. A possible explanation is that, during the process of filtration, calcium oxide inclusions (the products of deoxidation) are held on the surface of the filter (Figure 3). Only “active” calcium flows through into the mould.

The alumino-silicates seldom encountered in melts under normal circumstances are oval in shape and angular in character. After filtration, however, these inclusions took on a spherical form (Figure 7). At significantly low calcium content, therefore, an improvement in the shape of the non-metallic inclusions identified in the castings was evident. This effect may be explained by the fact that filtration significantly reduces the number of indigenous inclusions, whilst reductions in the turbulence of the flow following filtration reduces secondary oxidation in the mould. It follows that, instead of undergoing repeat deoxidation in the mould cavity, active calcium remains in solution and assists in the formation of globules.

It has been established that the number of oxide and sand inclusions getting through from the gating system (Figures 4 and 6) is reduced when using the filters in comparison with the use of a pressed filter (Figure 5). It is possible that the purging of non-metallic inclusions from the melt is linked with the large surface area of the filter.

Elongated sulphide inclusions were only encountered in one test case (Figure 8).

Figure 2. STELEX PrO fi lters

Figure 3. Inclusions trapped by the fi lter (oxides / slag)

Figure 4. Sectional view of the fi lter in the gating system

10

Globule-shaped non-metallic oxides were found dispersed in 15 out of the 100 photographic field views taken of the templates cut from the test castings (Figure 6). As regards the number and distribution of non-metallic inclusions and in comparison with melts under normal conditions, the application of filters appears to improve the quality of the steel.

A comparison of the earlier metallo-graphic analyses of melts under normal conditions with the test melts poured in the gating system fitted with the STELEX PrO filters, showed that use of the filters brought about an increase in non-metallic inclusion removal of 35 % (Figures 6, 7 and 8). Conclusions

1. The application of STELEX PrO filters is shown to be an effective means of combating melt flow turbulence in gating systems and inhibiting the processes leading to mould erosion. As a consequence, the incidence of slag and sand inclusions in the resulting castings is also reduced.

2. The use of filters in gating systems brings about reductions in the number of non-metallic inclusions caused by secondary oxidation in the mould at the time of pouring of the metal.

3. Steel castings poured under application of STELEX PrO filters demonstrate increased levels of purity, due to the removal of non-metallic sulphide, oxide and alumino-silicate inclusions.

4. Filtration of the steel promotes the formation of globular-shaped alumino-silicates which, in normal castings, are mostly encountered in an elongated, elliptical form.

5. The effective introduction of filtration may require multi-stage preparation including the development of specially-constructed gating systems. Improved melt and pouring procedures may compliment the performance of filters

References1. Svoboda J.M. et al. Trans. AFS 95 187-202 (1987)

2. Svalov N.V. “The filtration of large steel castings”, report to “Foundry Concilium No.5”, 2011.

3. Brown, John R. “Foseco Ferrous Foundryman’s Handbook”, 2000

Figure 5.

Figure 7. Spherically-shaped alumino-silicates (x 100)

Figure 8. Sulphide inclusions (x 100)

The use of fi ltration to improve the quality of large steel railroad castings

Figure 6. Globule-shaped oxides (x 100)

C Si Mn P S Cr Ni Al Cu V Ca

Melt 0.2 0.49 1.11 0.015 0.015 0.08 0.09 0.03 0.12 0.08 0.0013

Casting 0.19 0.49 1.09 0.016 0.011 0.08 0.09 0.03 0.12 0.08 0.0003

Melt 0.22 0.34 1.17 0.014 0.014 0.15 0.1 0.03 0.14 0.07 0.0009

Casting 0.2 0.35 1.13 0.014 0.013 0.15 0.1 0.03 0.14 0.07 0.0001

Melt 0.2 0.3 1.04 0.016 0.014 0.13 0.1 0.04 0.16 0.07 0.0011

Casting 0.18 0.3 1.06 0.017 0.013 0.13 0.1 0.04 0.16 0.07 0.0003

Table 1. Comparison of the chemical content of 20 GFL steel prior to pouring with the chemical content of the templates cut from the castings

Foseco International LimitedP.O. Box 5516

TamworthStaffordshire

England B78 3XQRegistered in England No. 468147

ISSN 0266 9994

Produced in England

400011330

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system of any nature or transmitted in anyform or by any means, including photocopying and recording, without the written permission of the copyright holder.

All statements, information and data contained herein are published as a guide and although believed to be accurate and reliable(having regard to the manufacturer’s practical experience) neither the manufacturer, licensor, seller nor publisher represents or

warrants, expressly or impliedly:

(1) their accuracy/reliability(2) that the use of the product(s) will not infringe third party rights

(3) that no further safety measures are required to meet local legislation

The seller is not authorised to make representations nor contract on behalf of the manufacturer/licensor. All sales by themanufacturer/seller are based on their respective conditions of sale available on request.

*FOSECO, the logo, SERT Metal, ISOMOL, STELEX and UCERAM are Trade Marks of the Vesuvius Group, registered in certain countries, used under licence.

® MAGMA is a registered Trade Mark of MAGMA Giessereitechnologie GmbH.

© Foseco International Ltd. 2012

COMMENTEditorial policy is to highlight the latest Foseco products and technical developments.

However, because of their newness, some developments may not be immediately available in your area.Your local Foseco company or agent will be pleased to advise.