New Methods During Development and Validation of Turbine Materials

0 downloads 136 Views 1MB Size Report
FEM computations and the analysis of turbine housings that were already .... homogenous as the FEM analysis software ass
Dipl. Ing. (FH) Frank Scherrer Dipl. Ing. Gerald Schall Dipl. Ing. Holger Gabriel Dipl. Ing. Steffen Bereswill

New Methods During Development and Validation of Turbine Materials (2008)

Introduction In the course of efforts to significantly lower the CO2 emission of the automobile sector, downsizing of motors has become a tried and tested means to the end. The development of such motors featuring smaller displacements will be driven yet further by most automobile manufacturers. If customer acceptance of these concepts is to be assured in all segments, then points such as agility, driving comfort and drivability must not be negatively influenced. This means that the specific output of motors must rise as well as the torque at lower engine speeds needing to increase. A consequence of this raising of the intermediate pressure level of motors is a rise in exhaust gas temperatures and therewith, of course, increasing thermal loading of the exhaust gas transporting components. The materials for turbine housings suited to exhaust gas temperatures greater than 820 °C can generally attribute their high temperature stability to their nickel content. The price for the alloying additive element nickel began to increase at the beginning of 2004 due to rising demand on the world market. June 2006 was marked by the beginning of an extreme pricehike that could not be explained alone by market demand dynamics. The background was speculative actions on the stock exchanges which were subsequently stopped by a rule change in June 2007 on the London Metal Exchange (LME) (Figure 1).

In the case of Borg Warner Turbo & Emission Systems, the cost pressures resulting from the nickel price development has led to investigations of new materials that can be used as 1

alternatives for D5S (35% nickel component). On the one hand, the durability of a turbine housing is dependent on the thermal load and, on the other hand, on the geometrical design. The standard material D5S is used hereby exactly in the same way for diesel applications with an exhaust temperature of only 820°C as it is for gasoline engine applications with an exhaust temperature of 950°C. Such usage results in a great need of a material whose high nickel percentage can lead to an incalculable cost development. Just like D5S, the standard material 1.4849 also contains over 30% nickel. For either of these materials and their application areas, it is advantageous if alternative materials could be investigated rapidly and economically.

Turbine housings and their characteristics The influence of geometrical design The customer normally specifies the installation space for the exhaust gas turbocharger (TC). This involves the specification of the orientation and position of the inlet and outlet flanges of a turbine housing. In addition, the connection to the manifold or cylinder head, respectively, the exhaust system is specified along with the flange shape and attachment method. Besides the geometrical guidelines, the motor data that is targeted as a goal is specified, which, in turn, leads to a required gas throughput and consequently to a spiral dimension for the turbine housing. The designer has the task of finding the best possible design for the turbine housing based on these specifications. Design and properties of the selected material lead to either the success or failure of the project, whereby design bears the essentially larger burden. In addition to thermal stressing of the turbine housing occurring due to internal strains, there are also mechanical constraining forces resulting from cycles of thermal expansion and contraction as well as vibrations. The two types of stress loads combine and may lead to cracks and deformations to the turbine housing. Material surface oxidation depends on the exhaust gas temperature, exhaust gas composition and the exposure time. Quite often, the design of the turbine housing can not be realized in such way that the superimposing forces do not overload the material. In this case, use of a higher quality material must be considered. If such scenarios involve diesel applications, D5S is often utilized. Normally, FEM computations are performed before turbine housings are cast using a different material. This is done to enable estimation of whether the material switch can possibly achieve the targeted goal, respectively, which design modifications are needed so that the material switch does indeed achieve the targeted goal. 2

Influence of material properties As already mentioned above, the cause of cracks and distortion of a turbine housing are strains of the component, which are due on one hand to thermal loading and on the other hand due to the occurrence of external mechanical influences. These load stresses lead to strains and possibly to the plastification of the material. If the material load is too great, the strain forces will be dissipated by creeping or crack formation. The interaction between the various material properties decides how high the strain forces will be. Cast iron materials are fundamentally differentiated by their shape of their graphite. The shape of the graphite (laminar, vermicular, globular) induces differing heat conductance capacities. The better the heat conductance, the lower the internal strain build-up within the component during heating and cooling. The chemical composition influences the yield strength, tensile strength, thermal expansion, specific heat capacity and modulus of elasticity. These material properties change according to the temperature. In order to sum up the advantages and disadvantages of a certain material, all properties must be considered in the light of transient temperature loading. To enable performance of a FEM computation, the material data must first be established. This procedure is both expensive and time-consuming. Moreover, the processing variations in the different foundries leads to the end result that materials behave variably despite possessing the identical chemical composition. The properties of a casting material are likewise not always absolutely identical and a FEM computation only represents an approximation of the reality.

Material spectrum and selection The rise in the price of nickel led to numerous attempts by many foundries to manufacture alternative materials with lower nickel content or improved material properties. This is true for both cast iron as well as steel casting. These materials, which include either new as well as "resurrected" materials, were offered to customers such as BorgWarner or to OEMs directly. Whether a new material is utilizable for an exhaust manifold or a turbine housing is decided either in advance by a FEM computation (if the material data is available) or, at the latest, during a test stand investigation. In view of the currently broad spectrum of options, a complete validation of all offered material variants on the test stand is not conceivable. The resulting costs and necessary time required for such would exceed any development budget. A preselection via a test that considers the critical characteristics of a turbine housing:

3

- cracking, - distortion, and - oxidation reduces the expenditure. Naturally, this test must be executed on geometrically identical test objects so as to balance out the influence presented by differing geometries.

Design of the test object Firstly, the critical points for the test of a turbine housing are to be discussed. Cracks appear predominantly in the areas of the turbine housings where walls of significantly varying thicknesses are adjacent to one another. FEM computations and the analysis of turbine housings that were already subjected to durability testing showed cracks at these places (Figure 2).

Moreover, deformations can be found on some turbine housings that are the result of creeping (Figure 3).

4

Crack formation and deformation are typical consequence of thermal loading of the material, respectively, of an alloy. It is therefore logical to construct the test object in such a manner that both effects can be made apparent. The geometry of the test object should be characterized by an extreme jump in wall thicknesses so as to provoke a crack in the transition area and it must exhibit zones that measurably deform upon application of a certain temperature for a certain time. Surface oxidation, that is, the formation of the scale coating, is yet another material-specific effect (Figure 4-6).

5

As the comparison of Figures 4, 5 and 6 shows, the scale coating can vary in its formation. However, the scale coating formation is not primarily dependent on the geometry, but rather on the heat admission, exposure time and composition of the atmosphere. Since the material tests generally used up to now were based on comparing interpretations (graphite formation, etc.), the test object should enable evaluation of material behavior on the basis of measurements and thereby rule out interpretations to a great extent. The number of cracks, lengths of cracks and flat surfaces that facilitate measurement before and after the test run are the fundamentals for a subsequent evaluation of the material. 6

Definition of the test object First of all, the test object should be able to be cast along with serial production parts in the running process. A laboratory manufacture is not desirable. Cores are to be avoided in order to exclude unnecessary costs and time factors. Furthermore, it should indicate the following effects in a representative test: - cracking - deformation - oxidation These requirements are fulfilled by the test object design shown in Figure 7:

The test object is characterized by the following details that are required to create or amplify the effects: - Wall thickness jumps to amplify strains during heating and cooling phases - Machined surfaces whose flatness is measurable - Small machining radius in order to provoke cracking - Continually increasing wall thicknesses should direct the gas flow into the test set-up

7

- Slight asymmetry should promote irregular build-up of strain

Already known materials are to serve as the basis for the investigation. Such materials have been used for years in the serial production of turbine housings for turbochargers. These test objects are to be heated at a specified gas temperature until the material in almost all areas of the test object reaches the gas temperature. This heating period is to be held constant for all materials. The series of heating and cooling phases with rapid switchover times causes an extreme thermal loading of the material, which leads to cracking and/or deformation of the test object in a relatively short span of time. After test is concluded, a microsection is to be prepared for investigation of the scale coating.

Definition of the test set-up The test object should be placed in a gas flow that attains to the greatest possible extent a unilateral heating of the test object. As a result of this, deformation of the test object's geometry should be achieved. Simultaneous employment of 2 test objects can be carried out to test two different alloys. When such simultaneous testing is performed, the gas flow should be arranged so that unilateral heating occurs to the greatest possible extent (Figure 8).

8

The hot gas is to be introduced in such a way that the main flow and thereby the principal component of the heat energy is applied to the test object unilaterally. Figure 9 shows a CFD analysis which emphasizes the heat distribution of the gas flow. In the upper portion of the Figure, one can recognize how the gas flow is introduced through both test objects. The test object holder clamps both test objects together with a spacer in between, leading to gas swirling. As the figure indicates, heat distribution varies within the set-up, which, in turn, promotes strain build-up. The hot gas is to be introduced in such a way that the main flow and thereby the principal component of the heat energy is applied to the test object unilaterally. Figure 9 shows a CFD analysis which emphasizes the heat distribution of the gas flow. In the upper portion of the Figure, one can recognize how the gas flow is introduced through both test objects. The test object holder clamps both test objects together with a spacer in between, leading to gas swirling. As the figure indicates, heat distribution varies within the set-up, which, in turn, promotes strain build-up.

9

10

FEM analysis of the test object The heating effect produces internal strain forces in the test object which the material, in turn, dissipates by cracking or creeping. Depending on the material type, chemical composition and the preexisting strain forces in the test object (resulting from the casting process and subsequent cooling), the test object reacts variably to the thermal loading. Figure 10 shows the heat distribution determined in the FEM analysis in the first 30 seconds of the heating process of the test object.

11

The variable temperature distribution shown in Figure 10 result in strain forces in the test object that lead to plastification. Figure 11 shows how extremely variably the material is stressed.

12

13

Evaluation of the test method The deformation ratio – that is the difference between flat portions of machined surfaces and roundness of the test object – as well as the number and length of cracks can be evaluated at the test's end. Since components that are manufactured from casting materials are never as homogenous as the FEM analysis software assumes, there may be small deviations from slab to slab. If a crack occurs, then the level of strain forces in the test object changes and, according to FEM, no further cracks are to be expected in the vicinity of the present crack. Since the outlined test method only concerns itself with a first preliminary test, a "general trend" result from 2 test objects suffices. As already mentioned above, the test object should be evaluated in regards to crack position and length as well as changes in flatness and roundness. The definition of some details must be included in this. Before the test, the test object is first to be marked with a heat resistant marking on one spot. This takes place on one of the two outer side surfaces, which can serve as a vent connection (Figure 13). The orientation of the marking defines further designation of crack positions and surfaces.

14

Surface ("S") flatness S1 to S4 are measured after processing of the raw part. After the test is carried out, the surfaces are once again measured for flatness. The differences between the measured values correspond to the deformation of the test object and can be represented by plotting in a diagram for comparison purposes.

15

Amerika

Europa

Asien

BorgWarner Turbo Systems 3800 Automation Avenue Auburn Hills, MI 48326/USA Telefon: ++ 1248 754 9600 Telefax: ++ 1248 754 9397

BorgWarner Turbo Systems GmbH Marnheimer Straße 85/87 67292 Kirchheimbolanden/Germany Telefon: ++ 49 (0)63 52 4 03-0 Telefax: ++ 49 (0)63 52 4 03-18 66

BorgWarner Turbo Systems P.O. Box 15075 Asheville, NC 28813/USA Telefon: ++ 1828 684 40 00 Telefax: ++ 1828 684 41 14

BorgWarner Ltd. Turbo Systems Division Bradford BD4 6SE West Yorkshire/UK Telefon: ++ 44 12 74 68 49 15 Telefax: ++ 44 12 74 68 96 71

BorgWarner Turbo & Emissions Systems Shin-Yokohama Turbo Branch Sumitomo Fudosan Shin-Yokohama Bldg.10F 2-5-5, Shin-Yokoyama, Kouhoku-ku Yokohama-shi, Kanagawa 222-0033 Japan Telefon: ++ 8145 470 6850 Telefax: ++ 8145 470 6811

BorgWarner Brasil Ltda. Estrada da Rhodia Km 15 P.O. Box 6540 13084-970 Campinas-SP/Brasil Telefon: ++ 55 19 37 87 57 00 Telefax: ++ 55 19 37 87 57 01

BorgWarner Automotive Components (Ningbo) Co., Ltd. Turbo & Emissions Systems No.188, Jingu Zhong Rd. (West) Yinzhou District Ningbo P.R. China 315104 Telefon: ++ 86 (0)574-8819-0930 Telefax: ++ 86 (0)574-8302-5883

BorgWarner France SAS 23 bis, avenue de l’Europe 78400 Chatou/Frankreich Telefon: ++ 33 1 3015 94 44 Telefax: ++ 33 1 3015 94 40

SeohanWarner Turbo Systems Ltd. Eoyeon-Hansan Industrial Park 893-1, Eoyeon-ri, Cheongbuk-myeon Pyongtaek-shi, Kyonggi-do 451-833 Korea Telefon: ++ 82 (0)31-680-00 01 Telefax: ++ 83 (0)31-683-929

BorgWarner Turbo Systems Poland Sp. z o.o. Jasionka 950 36-002 Rzeszów/Poland Telefon: ++ 48 1 78 50 88 00 Telefax: ++ 48 1 78 50 88 34

BorgWarner Turbo Systems Kft. Táncsics Mihály út 111. H-2840 Oroszlány/Ungarn Telefon: ++ 36 34 562-300 Telefax: ++ 36 34 562-301

www.borgwarner.com

feel good about driving