DEVELOPMENT Materials

32 MTZ worldwide 9/2002 Volume 63

Polymers are exposed to especially high loads in the enginecompartment of motor vehicles. For example, intake manifoldsmade of polyamide have to withstand a strong pressure peakduring engine backfires. Its bursting strength is therefore anessential item on OEM design specifications. In order to investi-gate the static and dynamic behaviour of the bursting strengthof friction-welded Ultramid intake manifolds, BASF developed atesting rig on which a series of experiments was carried outwith 80 series production manifolds.

1 Special Loads on the Manifold

The advantages of plastic over metal compo-nents are to be found in weight savings, cost-effective manufacture and almost unlimiteddesign freedom – factors that have accelerat-ed the use of automotive plastics in general,and specifically of polyamides in vehicle in-take systems [1,2]. The breakthrough for theplastic manifold came in the 1980s with the

development of a suitable manufacturingprocess – a fusible-core (or lost-core) tech-nique – which enabled an otherwise impossi-ble part to be moulded as a single piece.While fusible-core technology produces ex-cellent quality parts, the process is relativeexpensive – one reason why such units aremainly found on up-market models.

To ensure that the advantages of plasticswould not be lost on other classes of vehicle,

By Werner Wilhelm Kraft

and Hans-Peter Beringer

Berstdruckfestigkeit

reibgeschweißter

Polyamid-Saugrohre

You will find the figures mentioned in this article in the German issue of MTZ 9/2002 beginning on page 738.

Bursting Strength of Vibration-WeldedPolyamide IntakeManifolds

More than 30 years ago, BASF was one of thepioneers in the introduction of intake manifoldsmade of plastic. But it is nevertheless surprising

that Ultramid, the polyamide developed byBASF, is still used in this application even

today. The latest member of the polyamidefamily is called Ultramid HP, and is able to

withstand the high temperatures of the latestengine generations.

33MTZ worldwide 9/2002 Volume 63

where the cast metal manifold still dominat-ed, cheaper methods were developed in the1990s to make plastic units by welding to-gether two or more conventional injectionmouldings. An added advantage was thehigher degree of “functional integration”made possible by welded manifolds. Today,60 % of plastic manifolds are of the weldedtype, and their quality – the strength of theweld joint in particular – is increasing all thetime.

However, there is still a danger that abackfire occurring in the intake system maycause the plastic manifold to rupture alongthe weld line, Figure 1. It is therefore not sur-prising that the bursting strength is a centralitem on design specifications. The followingdescribes a testing rig developed at BASF’sthermoplastics application developmentcentre for optimising the bursting strength ofwelded plastic manifolds. The rig can also beused for developing pressure release valvesthat protect the manifold from excess pres-sure. Thanks to its high mechanical strengthand heat resistance, Ultramid, a polyamidedeveloped by BASF, is an established materi-al for moulding intake manifolds.

2 Static and Dynamic BurstingProperties

Finite element methods (FEM) can be used inthe design and development phases of an in-take manifold to accurately predict its burst-ing pressure. The basis of a successful designof a manifold lies in optimising the layout ofthe mould tool in order to minimise thewarpage in the individual parts (which aresubsequently welded together). Computer-aided techniques such as mould-fill simula-tion and warpage analysis are used withgreat success in meeting this aim. A furtherprerequisite for the accuracy of FEM lies inmatching the parameters of the friction-welding process to those of the material be-ing joined, Figure 2, enabling significant in-creases in weld strength to be achieved insome cases.

Until now, because of a lack of knowledgeabout the relationship between static anddynamic bursting strengths, it has not beenpossible to predict the dynamic burstingpressure with the same degree of accuracy.However, such knowledge is crucial to opti-mising the part, Figure 3.

In the past, the dynamic bursting strengthwas determined experimentally using a lim-ited number of pre-production units, whichwere often at different stages of develop-ment. Unfortunately, the number of experi-ments carried out was too few for statistical-ly significant results to be obtained, the ob-served values being too scattered for anyrecognisable differences between static and

dynamic bursting pressures to be deter-mined.

The results of an experiment carried outby BASF to determine the dynamic pressureof a statistically representative sample of 80production manifolds will be presented here.Along with other results, they show the rela-tionship between static and dynamic burst-ing pressures.

3 Description of the Test Equipment for TransientPressures on Parts

BASF has developed test equipment, Figure 4,capable of delivering pressures ranging fromquasi-static to 1 bar/ms (at the greatest dis-tance from the inlet valve). The equipmentessentially consists of a compressed air reser-voir, which can operate at up to 200 bar, anda computer-controlled high-speed valve. Thevolume of the reservoir can be varied, usuallybetween 0.5 and 5 litres. Sensors measure thevariation in pressure over time. The equip-ment can be operated in various modes: sin-gle step, pressure ramp, and cyclical.

4 Experimental Parameters and their Results

Both static and dynamic tests were carriedout using air as the pressure medium. To de-termine the effect of the pressure medium,six manifolds were made to burst using wa-ter as the pressure medium. The followingparameters were varied: type of loading (sin-gle step and stepwise), moisture content (dryvs. equilibrium moisture content), rate ofloading (static and dynamic), temperature,reservoir volume, and pressure gradient.

At the start of the trial, a statistical analy-sis based on 95 % confidence limits was car-ried out on up to six manifolds. As the trialprogressed, the values observed were foundto exhibit little scattering. Because of this, itwas possible for the remainder of the trial toreduce the number of parts tested for eachchange of parameter. The following observa-tions were made:■ Pressure medium: The type of pressuremedium used (water vs. air) had no effect onthe bursting pressure at room temperature(five parts tested with water, two parts test-ed with air). Parts tested with air exhibitedmore damage than parts tested with water.■ Type of loading: Dynamic single-step load-ing (up to 1 bar/ms) and step-wise loading(0.5-bar steps) produced the same level ofbursting pressure.■ Moisture content: The static bursting pres-sure measured on conditioned parts (labora-tory atmosphere with approximately 1 % rel-ative humidity) was 26 % higher than thatfor dry parts (six parts tested in each case).

■ Rate of loading and temperature: Thebursting pressure at loading rates of up to 1bar/ms was found to be 12 % lower than forquasi-static loading (six parts). Comparedwith room temperature, bursting pressurevalues at –40 and +120 °C fell by 21 % and 11 %respectively given the same rate of loading,and by 31 % and 23 % for quasi-static loading.Tests at loading rates of 0.2 and 0.4 bar/msrevealed a similar tendency.■ Reservoir volume: Increasing the reservoirvolume (between 0.5 and 5.0 litres) had no ef-fect on the bursting pressure, but it did causegreater damage to the part.■ Pressure gradient: Loading rates of morethan 1 bar/ms generated pressure wavessuch that the pressure profiles recorded ateach of the three sensors exhibited phase dif-ferences. The bursting pressure at a dynamicloading rate of 4 bar/ms (two parts) was 14 %higher than under a static load (six parts).

5 Summary

The investigation revealed that the burstingpressure of the intake manifold is indepen-dent of the pressure medium, type of loading(single-step or step-wise) and the volume ofthe air reservoir. Some parameters do how-ever affect the bursting strength: moisture-conditioned parts have a much higher burst-ing strength than dry parts; falls of up to 30 %in the dynamic bursting strength wererecorded for loading rates up to 1bar/ms,while pressure gradients above 1 bar/ms ledto higher bursting pressures. The results of aninvestigation carried out recently on anotherdesign of manifold, as well as the results ofmany individual – not statistically based –tests, revealed that, for loading rates of up to 1 bar/ms, the static and dynamic burstingpressures are comparable. It can therefore besaid, based on all the investigations carriedout to date, that for loading rates of up to 1 bar/ms, the difference between an intakemanifold’s static and dynamic bursting pres-sures depends on its design and can vary byup to 30 %. The investigations prove yetagain the need for dynamic testing and ex-perimental optimisation in the developmentof plastic parts, Figure 3 and Figure 5, in orderto identify design weaknesses or even avoidover-engineering.

References

[1] Fischer, Klaus W.: The World‘s First Thermo-plastic Intake Module Systems in BMW En-gines. New Powertrain Materials Processes,GPC 1998

[2] Schnell, S.: Wegbereiter zur Systemintegration,Automobil-Produktion, August 2000

[3] Kraft, W. W.: Untersuchung zweidimensionalerelastodynamischer Spannungskonzentrations-und Rissprobleme. Dissertation, UniversitätKaiserslautern, 1986

[4] Kraft, W. W.: BASF-interne Untersuchungs-Berichte, 1995–2001

MATERIALS Plastics