Foseco Ferrous Foundryman''''s Handbook Part 15 pdf

The six basic foundry cost elements are: Cost of molten metal in the mould currency unit/weight unit This includes costs of raw materials, energy, direct and indirect labour,melting depa

340 Foseco Ferrous Foundryman’s Handbook

The safety margin required

Feed demand from lower section

Superheated core or mould section

Once all the required inputs have been entered into the left hand side of theFeeder Size Calculation screen, the user may select a feeding product fromthe ‘Selected Product’ pull down list Once a product has been selected, theFEEDERCALC program calculates the optimum feeder size for the castingunder consideration and this information is displayed beneath the productname The next size up in the product range is also displayed for comparisonpurposes

The Side Neck Calculation button allows the user to calculate the size of

a neck used with a side feeder The neck modulus, provided by the feedersize calculation for the feeder to be used, is automatically transferred to thisscreen The program then calculates the range of values which one dimension

of a rectangular-section neck would have based on this neck modulus Whenthe most convenient dimension for the casting being considered is entered,the second dimension is automatically calculated together with the resultingcontact area

alternative feeding practices can also be made (for example a practice usingsleeves compared against one using sand risers).

By clicking on the ‘Cost Analysis’ tab, the screen Fig.19.33 appears Sixitems of basic foundry cost data are presented These are average valuesestimated by Foseco (for the country for which the particular program wasdesigned) The operator may simply accept these (and continue onto the JobCost button); or the operator may enter actual costs for the foundry andsave this data set by selecting the ‘New’ or ‘Save’ buttons

Figure 19.33 Cost analysis screen.

The six basic foundry cost elements are:

Cost of molten metal in the mould (currency unit/weight unit)

This includes costs of raw materials, energy, direct and indirect labour,melting department overheads and any penalty for losses in providingliquid metal to the mould cavity, i.e melting and pouring

Value accorded to returns (currency unit/weight unit)

The value allocated to the metal in feeders, running systems, and scrapcastings etc which are returned for re-melting

Cost of cutting (feeder removal) (currency unit/area)

This includes costs of raw materials, energy, direct and indirect labour,cleaning department overheads and penalty for losses in cutting feedersfrom castings

Cost of grinding (feeder removal) (currency unit/area)

This includes costs of raw materials, energy, direct and indirect labour,

342 Foseco Ferrous Foundryman’s Handbook

cleaning department overheads and penalty for losses for grinding thefeeder stub etc in cleaning castings

Sand density (weight unit/volume unit)

The density of the moulding medium (sand) when compacted in themould (typically 1.5 g/cm2 for silica sand)

Cost of moulding sand (currency unit/weight unit)

This includes the costs of raw materials, energy, direct and indirectlabour and overheads associated with preparing the moulding sand(i.e sand/binder storage, transport and mixing)

Job costs

The Job Costs button accesses the screen of Fig 19.34 on which details of thecasting and feeding are entered allowing two practices to be compared Thefinal Report screen, Fig.19.35, lists all the costs entered and calculated withthe total costs for each practice displayed Various ‘what if?’ scenarios may

be investigated by changing data on the Base Costs or Job Costs screen

Figure 19.34 Job costs screen.

Authorisation

The FEEDERCALC program uses a licensing system to operate the software

If an attempt is made to copy the files, or run the program without properauthorisation, the program will not run and an error message pertaining to

authorisation will be shown on the screen Authorisation to run the program

is obtained when the Serialization Code for the specific computer on whichthe program is to be run is notified to Foseco An Authorisation Code is thenprovided by Foseco enabling the program to run The program is set toexpire after one year The user must send Foseco a purchase order to receivethe new Authorisation Code

Figure 19.35 Report screen.

of filling a mould with liquid metal and its subsequent solidification could

be accurately and quickly modelled by computer, shrinkage cavities andother potential defects could be predicted The effect of changing the gatingsystem, the position and size of feeders and even the casting design could

be simulated The casting method could then be optimised before the designand method are finalised, so avoiding expensive and time consuming foundrytrials

Software for the numerical simulation of flow and solidification duringcasting processes has been available since the mid-1980s A large number ofcommercial software packages are now available and they are improvingall the time The modelling of heat flow and solidification of castings is nowwell advanced Modelling the filling of castings is more difficult since bothturbulent and quiescent flow in complex shaped cavities may be involved.The effects of surface oxide films and bubble entrainment are furthercomplications and it is not easy to check the predictions experimentally incomplex moulds

Solidification modelling

The aim of solidification modelling is to:

Predict the pattern of solidification, indicating where shrinkage cavitiesand associated defects may arise

Simulate solidification with the casting in various positions, so that theoptimum position may be selected

Calculate the volumes and weights of all the different materials in thesolid model

Provide a choice of quality levels, allowing for example the highlighting

or ignoring of micro-porosity

Perform over a range of metals, including steel, white iron, grey andductile iron and non-ferrous metals.

A number of systems are available, they may be divided into two basictypes:

Numerical heat flow simulations

Empirical rule based simulations

Numerical based systems, of which the best known is MAGMAsoft, arebased on thermophysical data: surface tension, specific heat, viscosity, latentheat, thermal conductivity and heat transfer coefficients of metal, mouldand core materials Mathematical equations are used to calculate heat flowand to predict temperatures within the cooling casting The complexities ofthe equations do not allow direct solutions so numerical methods of solvingdifferential equations, finite difference or finite element techniques, must beused Both require considerable computing power The time and positionwhere solidification commences is predicted and regions within the castingidentified which are likely to become isolated from feed metal Turbulenceduring mould filling and convection effects during cooling need to be takeninto account

Empirical rule based systems, such as SOLSTAR, take a model of thecasting and its surrounding moulding media divided into small cubicelements Heat flow and solidification are then modelled by applying iterativerules In the SOLSTAR method, each element is considered to be at thecentre of a Rubik cube of elements with 26 nearest neighbours The subsequentheat exchange calculations are then carried out in 26 directions taking intoaccount the temperatures and properties of the neighbouring sites Thetemperature of a particular site is thus changed step by step resulting in anaccurate prediction of thermal history A liquid site will transform into solidwhen the site reaches a predetermined value Metal flow from neighbouringelements is then mathematically simulated to take up the space vacated byshrinkage When there are no liquid metal elements left to fill the shrinkagecavity a void is created The system is thus able to predict where shrinkagedefects are likely to occur in the casting By ‘calibrating’ the rule basedsystem against experimental results, accurate prediction of shrinkage defects

Rule based systems use standard PC-based computers and are designed

to be used by an average foundry engineer Simulations of freezing of castingsare achieved in a fraction of the time needed by numerical models Numericalsystems require more computing power, needing a workstation costing severaltimes more than a PC and requiring a highly trained computer operator.Both systems are useful to the practical foundryman, not only cutting out

346 Foseco Ferrous Foundryman’s Handbook

trial and error sampling but increasing metal yield, reducing lead times,optimising production methods and improving the accuracy of quotations.The speed and user friendliness of rule based programs makes thembetter suited to jobbing foundries where many different castings need to beprocessed and computation time is an issue On the other hand, numericalprograms demand strong computing power and lengthy processing time,but as better physical data becomes available, they can in principle providegreater accuracy and a closer representation of what actually happens in themould They allow temperature profiles after solidification to be modelled

so that metallurgical structures can be predicted They may even take intoaccount the thermal effects that occur during the filling of the mould Currentlysuch programs are perhaps better suited to research and to study highlyengineered, critical components They are also used for important productdevelopment projects, for example automotive castings to be made in verylarge numbers

Mould filling simulation

Many of the problems associated with the casting process are related topoor mould filling Programs have been developed which allow mouldfilling to be simulated The aim is to predict how running and gating designaffects turbulence in the mould which may trap oxidation products andcause potential defects MAGMAsoft (and others) have developed suchprograms which allow the visualisation and animation of the movement ofthe melt surface during filling Current limitations are system time, it cantake several days to simulate the flow in a mould, and the lack of good data

on surface tension, viscosity etc Several laboratories around the world aregenerating the thermo-physical data needed to improve the simulations

The SOLSTAR solidification program

SOLSTAR is used by a large number of foundries, service bureaux andeducational establishments The breakdown of usage (in 1994) is:

Non-ferrous foundries 10%

Education & Service 27%

The procedures for carrying out a SOLSTAR analysis are:

(1) Using the casting drawing, determine model scale and element size.(2) Make the solid model of the casting

(3) Make the solid model of the proposed production method (feeders,

chills, insulators etc.) Use the program’s own feeder-size calculator

if required

(4) Carry out thermal analysis to establish the order of solidification.(5) Carry out solidification simulation to a set quality standard, for theselected alloy incorporating shrinkage percentage, ingate effects etc.This results in the model being changed to the predicted final shape(internal and external) of the casting showing size, shape and location

of shrinkage cavities in casting and feeders

(6) Examine the predicted shrinkage (the equivalent of non-destructivetesting) by viewing and plotting of 3D ‘X-rays’ and sections of themodel in 2D slices or 3D sections and relating predicted defects tosolidification contours and required quality standards

(7) If the predicted defects do not meet the required quality standard,develop an improved production method and repeat the procedures.These trial-and-error sampling procedures can be carried out very rapidly,allowing the operator to indulge in any number of ‘what-if’ experiments

Solid modelling

The first stage in any solidification simulation is to create a three-dimensionalmodel of the component with its associated method This will often take thegreatest proportion of time, as much as 70% The SOLSTAR program has itsown solid modeller/mesh generator capable of modelling the most complexcasting shapes Depending on the computer hardware specification, modelscan contain up to 256 million elements but most models use between 2 and

64 million elements Figure 20.1 shows a solid model of a 350 kg steel valvecasting containing 40 million elements produced in less than 3 hours

It is possible to transfer 3D models from any other CAD system usingStereolithography STL files created by them These models can then bemanipulated within the solidification software so that the method can beadded

Thermal analysis

The thermal analysis calculates the simulated heat flow between the elements

of the solid model which gives a ‘thermal picture’ of the conditions prevailing

at a specific point in time SOLSTAR’s thermal analysis simulates ‘heatflow’ in 26 directions, with each cuboid element of metal, mould, chill etc.being the equivalent of the centre block of a ‘Rubik’ cube (27 cubes).SOLSTAR uses the thermal analysis to store details of the solidificationorder of each element of the casting and feeding system Figure 20.2 shows

a ‘thermal’ illustration of a section through the steel valve casting (in blackand white on p 349 and reproduced in colour in plate section) This isproduced in colours showing ‘solidification contours’ of the metal from the

348 Foseco Ferrous Foundryman’s Handbook

Figure 20.1 Solid model of a valve casting utilising 40 million elements (This figure is reproduced in colour in plate section.)

Figure 20.2 Solidification contours for the lower section of the valve casting (This figure is reproduced in colour in plate section.)

end of pouring to the end of solidification The thermal analysis for the steelvalve involves approximately 200 billion ‘heat exchanges’ between adjacentelements of the model and was calculated in 22 minutes using a 266 MHzcomputer

Solidification simulation

After the thermal analysis is completed, each metal element of the model isallocated an order of solidification SOLSTAR then carries out a solidificationsimulation of the metal elements, by solidifying them in the predeterminedorder During this simulation several things are happening:

(1) The solidified elements are assumed to have increased in density,accompanied by a loss of volume

(2) This loss (liquid shrinkage) is calculated by multiplying the number

of solidifying elements by the input shrinkage factor for the alloy.(3) The software calculates (according to the alloy and the required qualitystandard) whether this shrinkage will manifest itself in the form of acavity and, if so, how big the cavity will be

(4) The resultant cavity is placed in the remaining liquid of the section ofwhich it is a part Where it resides in this remaining liquid depends

on the type of alloy

350 Foseco Ferrous Foundryman’s Handbook

(5) The program continually checks the linkages between the remainingliquid metal

(6) Metal elements are continually ‘flowing’ through the liquid paths toreplace volume loss during solidification, so accurate tracing of thesepaths is critical to the program

At the end of the solidification simulation the model represents the casting(and feeders) at the ‘shake-out’ stage of production Figure 20.3 is an ‘X-ray’plot of the final model showing all the shrinkage cavities predicted to beoutside of the requested quality standard A Class 2 solidification simulationfor this valve took 26 minutes using a 266MHz computer

Figure 20.3 An ‘X-ray’ plot showing predicted shrinkage cavities (This figure is reproduced in colour in plate section.)

During the solidification simulation, the effect of varying

moulding position,

ingate position,

mould materials, chills, insulating and exothermic materials,

can be modelled, allowing the optimum method of making the casting to bepredicted