R E S E A R C H CYLINDER HEAD
AUTHORS
DIPL.-ING. (FH) MARC LEIDENFROST
is Research Engineer at the Institute of Materials Science and Mechanics of Materials of the Technical University Munich (Germany).
PROFESSOR DR. MONT. MONT. HABIL. DR. RER. NAT. H. C. EWALD WERNER
is Head of the Institute of Materials Science and Mechanics of Materials and of the State Material Testing Laboratory in Mechanical Engineering of the Technical University Munich (Germany).
70
MEASURING TEMPERATURE AT A CYLINDER HEAD UNDER THERMAL LOAD Before the lifetime of thermo-mechanical stressed cylinder heads can be evaluated numerically, the temperature field at the cylinder head must be known from a thermal analysis. This temperature field has an essential influence on the predicted lifetime of the cylinder head. Hence, a comparison between measured and calculated temperatures is necessary. In this paper a method method for the measuremen measurementt of cylinder cylinder head temperatures, which was investigated at TU Munich, is presented. The thermocouples are applied in boreholes that are inserted in parallel to the flame deck. They are fixed into the boreholes with a high-temperature cement.
1
INTRODUCTION
2
TEMPERATURE MEASUREMENT
3
DRILLED HOLES AND ADHESIVE FOR SENSORS
4
SENSOR POSITION
5
SUMMARY
1 INTRODUCTION
The lifetime of cylinder heads in modern diesel engines for commercial vehicles is determined primarily by mechanical loads (such as ignition pressure) and cyclic thermal loads (such as start/stop operation). For reasons of time and costs, increased use is made of the findings from finite element (FE) s tudies in lifetime assessment. Particularly in terms of modelling the effects of thermo-mechanical stress and strain on the lifetime of cylinder cylinder heads, temperature distribution, temperature-variation temperature-variation span, maximum temperature in the cylinder head and temperature/time response are decisive. Computed temperature distributions have to be verified in advance on the basis of measured values to permit viable assessment of the results of simulation. In the main, these measured values are obtained from component tests or engine tests. Because of the principle involved, the highest and therefore the most significant temperatures in the cylinder head occur at the surface inside the combustion chambers. Ideally, methods can be employed to measure these temperatures directly without influencing them. Under all other circumstances the possible influencing factors and their effects on the measured values must be known or obtained. Given the high mechanical and thermal loads associated with the combustion proces s in commercial-vehicle commercial-vehicle diesel engines, direct measurement inside the combustion chamber or at the piston is subject to certain constraints. Similarly, Similarly, soot deposits preclude the cost-effective use of methods with optical access to the combustion chamber. One practical way of measuring temperatures in close proximity to the combustion-chamber roof is to drill the cylincylinder head and insert thermocouples. This method is discussed by the temperature-measurement results obtained in a component test with three individual-cylinder heads from MAN Truck & Bus AG [1]. The report focuses primarily primarily on the results of the studies with regard to the influence of sensor position on measured values. The effect on the results of the drilled holes and the adhesive used is also examined. Regarding the individual points, FEM calculations were undertaken for the purposes of closer analysis. The results obtained in this way are discussed at the appropriate point.
❶ Test bench setup (Ch = cylinder he ad)
tures individually by means of the three gas burners. The cylinder heads used for the purposes of this test are MAN Truck & Bus AG individual-cylinder individual-cylinder heads made of cast iron with vermicular graphite. The cylinder heads used were pre-grouped in accordance with series production and bolted to the adapter plate. Three measuring cylinder heads fitted with thermocouples were mounted on the test bed for the purp oses of temperature measurement. In equivalence to temperature measurements in past engine tests, the measuring cylinder heads had 16 temperature measuring points. The holes (diameter: d = = 3 mm) for the thermocouples were drilled 5 mm from the combustion-chamber roof. ❷ shows the CAD model of the measuring cylinder with a section through the holes for the thermocouples, drilled parallel to the combustion-chamber roof. Two thermocouples were installed in each valve brigde, where the temperatures are hottest. A sp ecial high-temperature high-temperature adhesive was used to locate the thermocouples in the drilled holes. The test-bed parameters were adapted iteratively in such a way that the measured maximum temperatures at the end of the warmup phase correlated with measured values obtained from prior thermal-shock engine tests. Temperature measurement took place after a running-in phase so that initiation transients would not affect the results. The temperatures were recorded over a number of cycles.
2 TEMPERATURE MEASUREMENT
In the framework of initial studies of the lifetime lifetime of cas t-iron cylinder heads, the cyclic thermal loading occurring in thermal-shock engine testing was modelled in simplified terms. A cylinder-head test bed manufactured by IABG was used for this purpose [2, 3]. Temperature measurement was conducted to calibrate the test bed on the one hand and to verify the thermal simulation simulation on the other. The test-bed configuration, ❶, consists primarily of three individual-cylinder heads, an adapter plate (as a substitute for the crankcase) and one gas burner per cylinder head. The cylinder cylinder heads are interfaced to the recirculating coolant system via the adapter plate. Coolant supply was set up in such a way that coolant inlet mass flow to all three cylinder heads was virtually identical. Each cylinder head was heated and brought to specific combustion-chamber tempera07-08I2012
Volume 73
❷ Bore holes for the thermo-couples in a single cylinder head
71
R E S E A R C H CYLINDER HEAD
❹ shows the computed temperature distribution at the end of the
By way of example, ❸ shows a section of the measured temperature/time transients for measuring point No. 15 (exhaust-valve bridge) of the three cylinder heads. The measurement data obtained in this way constitute the basis for the necessary crosschecking between tes ting and modelling.
warm-up phase for one cylinder head without and one cylinder cylinder head with drilled drilled holes for the sensors. The figure shows the exhaust-valve bridge in a section normal to the combustion-chamber roof through measuring point No. 15. It is clear that in the cylinder head with holes drilled for the sensors the temperatures acting from the cylinder-head roof through to the hole are considerably higher. higher. This is due to the thermal conductivity of the high-temperature adhesive used, which is approximately 20 times less than that of the base metal of the cylinder head. This in turn means that the drilled drilled holes for the sensors act like a local insulating layer. This has an effect primarily on temperature distribution in the valve bridge, firstly because starting from the combustion chamber roof the temperature gradient here is very high, and secondly because the drilled holes extend across virtually the entire width of the valve bridge. When due provision for the drilled holes for the sensors is made in the FE calculations, the temperatures at the target positions of the thermocouples 5 mm below the combustion-chamber roof are up to approximately 4 % higher than in the counterpart FE calculations without these holes. With regard to subsequent life-time assessment, in which temperature distribution and temperature variation span play a crucial role, when measurement and simulation are cross-checked due allowance has to be made for the holes drilled for the sensors and for the adhesive used.
3 DRILLED DRILLED HOLES AND ADHESIVE FO R SENSORS SENSORS
4 SENSOR POSITION
At variance with the practice adopted for the temperature measurements, the cylinder heads used for the lifetime trials trials undertaken in component testing were not drilled to take measuring sensors. Consequently, this justifies asking whether the effect on the temperature field in the cylinder head of the holes drilled to take sensors and of the adhesive used has to be taken into account when temperature measurement is remodelled. To explore this in more detail, detail, the temperature field was calculated in two different FE calculations and the results were compar ed. One of the FE models allowed for drilled holes for sensors, equivalent to temperature measurement, measurement, and the other was structured without these holes. In both instances linear tetrahedron elements were employed for discretion. Calculation was transient because the results of temperature measurement, measurement, ③, sufficed to show that a stationary condition is not reached at the end of the heating and cooling phase. Sets of three thermal cycles were modelled, with the last cycle of each set being evaluated.
In chapter 3, for evaluation of the calculations it was assumed that the thermocouples are situated at the target positions defined in the CAD model. In the real part, by contrast, deviations between target position and actual position have to be reckoned with. Given the inhomogeneous temperature distribution and the high temperature gradient from combustion-chamber roof toward coolant galleries, ④, these differences in position are reflected in the measured temperatures. The resultant scatter in the measured values is of interest, par ticularly ticularly with regard to crosschecking the measured results and the results obtained by computation. There are two different causes for these differences between target position and actual position of the thermocouples. One is that because of its low stiffness and the long distance drilled, the relatively thin twist drill (d = = 3 mm) tends to drift off-line as it penetrates the cylinder-head material. This is unavoidable even despite the fact that the holes are drilled with state-of-the-art CNC machines. Another cause is the smaller diameter of the thermocouples vis-àvis hole diameter. This makes it impossible to guarantee the exact position of the sensor relative to the centreline of the drilled hole (target position). To illustrate this, the positions of the thermocouple tips in one of the measuring cylinder heads used for the test were ascertained. Firstly, the head was split 10 mm below the combustion-chamber roof. The sensors were located within the 10-mm thick slice of cylinder head obtained in this way. This slice was ground down from the w ater-chamber face until the sensors were visible in their drilled holes. The cylinder-head slice was X-rayed so that the differences in the positions of the sensors in the plane paralleling the combustion chamber could be measured. Evaluation Evaluation was undertaken using the X-ray images and reference points defined beforehand. The differences in position normal to the combustion-chamber roof were then ascertained. For this purpose the drilled holes with the sensors were sawn out of the cylinder-head slice; the saw cut was through the sensor tips, normal to the combustion-chamber
❸ Temperature/time-profiles of cylinder head temperatures, measured at measuring point No. 15 (exhaust-valve bridge)
❹ Calculated temperature field at the cylinder head without (left) and with bore holes (right) at the end of the heating phase
72
❺ Measured difference between desired and actual position of the thermo-couples
❻ Change in measured temperature due to position deviation of the thermo couple in axis-direction of the borehole
roof. The distance between combustion-chamber roof and the tip of each sensor could then be measured. In this way it was possible to ascertain the differences between target position and actual position in the direction normal to the combustion-chamber r oof. In order to ensure sufficient measuring accuracy, all measurements were carried out on a coordinate measuring machine. ❺ shows the deviations in all three spatial directions for the 16 measuring points. As the measurements show, in xy (plane paralleling the combustionchamber roof) the differences are in some instances considerable, with a variation of up to 2.5 mm. In z direction, in other words normal to the combustion-chamber roof, the maximum difference measured between target position and actual position of the sensors was 0.5 mm. Given the temperature distribution in the cylinder head illustrated in ④, the deviations in xy can be expected to have only a slight effect on the measured values. T his is due primarily to the slight temperature gradient in these coordinate directions. Conversely, the differences in z direction will will probably have a considerable effect on the measured values. The purpose of fur ther investigations was to qualify qualify the expected effects of the differences between target position and actual position of the sensors on the measured temperatures. The studies necessar y to this end were conducted on the basis of FE computations. To eliminate eliminate the need for generating a new CAD model for each sen sor position, the initial calculations were carried out using an FE model without drilled drilled holes for the sensors. The differences between target and actual positions considered in these studies were along and normal to the axis of the drilled hole and normal to the combustion-chamber roof. ❻, ❼ and ❽ show the temperature difference at the sensor tip as a function of positional variation. By way of example, in this case the measuring points evaluated were those with the highest measured temperatures (measuring points 14 and 15). ⑥ illustrates differences in position along the axis of the drilled drilled hole. The plot shows that at differences up to ±2.5 mm there are no significant differences in temperature at the measuring point. ⑦ shows the computed temperature deltas for positional differences normal to the axis of the drilled hole in the plane parallel to the combustion-chamber roof. As the diagram indicates, considerably higher temperature differences are measured at drilled holes closer to the exhaust duct. A positional difference of 2.5 mm toward the exhaust duct suffices to yield a temperature difference of 20 K. A shift toward the middle of the valve bridge produces negligible negligible differences in measured temperatures. The effect of the pronounced 07-08I2012
Volume 73
❼ Change in measured temperature due to position deviation of the thermocouple normal to the axis-direction of the borehole
❽ Change in measured temperature due to position deviation of the thermocouple normal to the flame deck
temperature gradient mentioned above is apparent in differences between sensor positions normal to the combustion-chamber roof, ⑧. A difference of ±0.5 mm between target position and actual position of the sensor produces a ±10 K difference in measured temperature at the measuring points in the hottest part of the cylinder head. The same effects were obser ved at the remaining measuring points in the “colder” parts of the cylinder head, although
73
R E S E A R C H CYLINDER HEAD
❾ Change in measured temperature due to position deviation of the thermocouple normal to the flame deck taking into account the boreholes
absolute differences are not as pronounced. In the main, this is due to the much less temperature gradients in these areas. Based on the results obtained from the first studies with the FE model without drilled holes for the sensors, the drilled holes for the thermocouples were taken into account in the FE model in a fur ther study. The positional deviations in the plane parallel to the combustion-chamber roof were not evaluated in this context, because the drilled drilled holes and the high-temperature adhesive primarily affect the temperature gradients in the vertical direction, ④. The results of the calculations without drilled holes for the sensors can be used to assess the differences. Normal to the combustion-chamber combustion-chamber roof, the thermocouples should be positioned 5 mm below the surface, in other words on the axis of each drilled hole. Because of the diameter dia meter of the thermocouples, which is smaller than that of the drilled hole, the actual positions are off the axis of the resp ective drilled drilled holes. Even without provision being made for the drilled holes, as already shown in ⑧, this results in temperature differences of ±10 K for a ±0.5 mm positional delta between the measuring positions. ❾ shows the calculated temperature difference as a function of the offset from target tar get position with the drilled holes for the sensors taken duly into account. As in the earlier results, the differences obs ervable in the measured temperatures are again large even when the difference in position is small. The temperature field is further influenced by the insulating effect of the drilled holes for the sensors and the high-temperature adhesive. Under worst-case conditions the thermocouples are in contact with the wall of the drilled holes, in which case they assume the minimum and maximum distances respectively from the combustion chamber. This leads to calculated calculated temperature differences between +70 K and -35 K at the measuring point.
perature adhesive employed to locate the thermocouples. The conclusions can be summarised as follows: : The temperatur temperature e field in the cyli cylinder nder head and the the measured measured values are influenced by the considerably lower thermal conductivity conductivity of the high-temperature adhesive vis-à-vis the bas e metal of the cylinder head. : With regard to the measured temperatures, deviations (up to ±1 mm) of the sensors from the target po sitions in the plane paralleling alleling the combustion-chamber roof play a negligible role. : Positional deviations normal to the combustion combustion chamber chamber have a very large effect on the measured temperatures. In this respect and depending on the measuring point, temperature differences up to +70 K and -35 K are possible. Temperatures in the cylinder head in a plane parallel to the combustion-chamber roof can be measured by the method described here. Conclusions about cylinder-head temperatures at the combustion-chamber roof can be drawn, with appropriate experience and knowledge of the components. As regards crosschecking these measured values with the re sults obtained by simulation, simulation, due provision has to be made for the effect of the adhe sive-filled drilled holes. The choice of high-temperature adhesive offers an approach to improving the method of measurement discussed here. Thermal conductivity closer to that of the cylinder-head material would reduce the adhesive’s effect on the temperature field. Further improvements can be achieved by positional accuracy, particularly normal to the combustion-chamber roof. REFERENCES [1] Leidenfrost, M.; Werner, E.; Meyer, D.: Lebensdauerbewertung von thermo-
mechanisch belasteten Gusseisen-Zylinderköpfen. In: Materials Testing 53 (2011), No. 11/12, pp. 758 – 762 [2] Achatz, F.; Fröschl, J.; Rödling, S.; Gärtner, J.: Innovatives Bauteilprüfkonzept für Zylinderköpfe. In: MTZ 70 (2009), No. 12, pp. 930 – 934 [3] Achatz, F.; Fröschl, J.; Rödling, S.: Zeitgeraffte und kostenoptimierte Bauteilprüfung an Zylinderköpfen. In: Materials Testing 52 (2010), No. 7/8, pp. 502 – 507
THANKS
5 SUMMARY
The calculated temperature field is crucial for numeric assessment of lifetime of thermo-mechanically stressed cylinder heads. Computed temperatures have to be verified by comparison with measured values from component or engine tests. This paper considers temperature measurement by thermocouples inserted into holes drilled drilled parallel to the combustion-chamber roof. In this context particular ticular attention is paid to the influence on the measured results of sensor position, the effects of the drilled holes and the high-tem-
74
The authors thank the MAN Truck & Bus AG for funding the studies and supplying the cylinder heads. Special thanks to Dr. Meyer, MAN Truck & Bus AG, for the acquisition of the project management and to the employees of the departments EMTS and EMTM (plant Nuremberg) for their support.
