DEVELOPMENT OF A LOW DRIFT TYPE K THERMOCOUPLE CABLE FOR AEROSPACE APPLICATIONS

Transcription

1 DEVELOPMENT OF A LOW DRIFT TYPE K THERMOCOUPLE CABLE FOR AEROSPACE APPLICATIONS Dr Michele Scervini Research Scientist University of Cambridge Trevor D Ford Technical Director CCPI Europe Limited ABSTRACT There is an increasing demand in both engine operation and in the manufacture of aerospace/aircraft parts and systems to measure temperature, particularly high temperatures, accurately. The existing base metal thermocouple designs used to conduct these measurements struggle to fulfil this critical measurement role due to instability and drift. This paper looks at why drift in base metal thermocouples occurs and how from this information a new design of mineral insulated thermocouple cable for sensor manufacture has been designed. The operational features and advantages of the new cable design will be discussed and results of comparative drift tests under cyclic temperature conditions will be demonstrated. The paper will discuss how the new design will offer the opportunity for type K and type N thermocouples to work for longer and at higher temperatures with significantly reduced drift giving increasing confidence in measurement capability. The new designs has raised significant interest in both scientific and industrial environments and independent testing by National Metrology Institutes for both high temperature and long duration has been planned as part of future research projects. INTRODUCTION Base metal thermocouples such as the type K (Nickel Chromium v Nickel Aluminium) sensors have been, and continue to be, the mainstay of manufacturing industry for temperature measurement. With the development of high performance materials for use in the aerospace and nuclear industries there is a growing demand to measure temperature to a greater accuracy and at higher temperatures. During the last decade the need for temperature sensors to work under these demanding conditions has increased to a point where the limitations of the type K thermocouple are being reached and in some cases even exceeded. The development of the type N (Nicrosil v Nisil) thermocouple and its application particularly for high temperature operation has allowed base metal thermocouples to continue to be used as the primary sensor for operations such as load monitoring and thermal uniformity surveys. However both the type K and type N thermocouples in operational conditions during extended and/or high temperature conditions have a limited operational life. This has led to manufacturers and heat treaters having to look at using the less versatile and significantly more expensive Platinum/Rhodium (type R, S and B) noble metal thermocouples in temperature ranges above 1100 C (2012 F) for reliable and accurate measurements. BASE METAL THERMOCOUPLE LIMITATIONS This set of conditions was the driving force that started a long term investigation in to understanding why base metal thermocouples drift by research scientists in the Department of Materials Science and Metallurgy at the University of Cambridge. Work was done looking at the root causes of why particularly the type K thermocouple tends to show accelerated drift under certain conditions and at high temperatures. Significant work has been conducted over the last 50 years into this field and most of the general mechanisms associated to drift were already known. The following mechanisms can be identified among the major causes of drift in base metal thermocouples: 1. Oxidation 2. Short range ordering 3. Depletion/change in composition 4. Other physical/chemical changes Oxidation was a major issue for bare wire thermocouples as it produced extensive change in composition as a 1

2 result of the formation of oxide scales on the surface of the thermoelements. The use of a sheath in the mineral insulated thermocouple cable configuration has had a positive effect on reducing the contributions of oxidation: this has been clear for many years and the current mineral insulated construction design for base metal thermocouples has been the major reason why the impact on drift of oxidation and other physical/chemical changes caused by the aggressive operating environment have been significantly reduced under operational conditions. The development of the type N thermocouple combination in the 1960 s had a positive effect on providing a temperature sensor with a reduced contribution of short range ordering, a condition that has a major contribution to drift in the lower temperature range. The factor that had not been addressed is the depletion/change in chemical combination of the conductors; in fact the mineral insulated construction has increased the level of depletion/change that occurs. Studies conducted by researchers at Cambridge have shown that base metal mineral insulated thermocouple conductors exposed at high temperatures experience increased levels of contamination from elements such as Mn and Cr. Also these studies have revealed that the levels of these elements had penetrated deep into the thermocouple conductors after limited operational use. The source of these and other chemical elements have been identified as being i) the other thermocouple conductor ii) the outer sheathing of the mineral insulated thermocouple cable. Further tests conducted at Cambridge have shown the major source in terms of both volume and material that causes higher rates of drift is the mineral insulated thermocouple outer sheathing. As would be expected, the higher the temperature of operation the greater the rates of these changes. This varying amount of changes down the thermocouple conductor s length results in an increase of the heterogeneous nature of the base metal thermocouple conductors, therefore having a significant contribution to the observed rate of drift. high temperature, the major source of changes in composition of the conductors and therefore source or cause of thermocouple drift. The same sheath that had allowed thermocouples to have longer operational life compared to bare wire sensors has become the major constrain to increase further the operating temperature of the sensor and to fully exploit the potential of base metal thermocouples at the highest temperatures. The work conducted by researchers at Cambridge focused not only on the mechanisms and root causes associated to drift but also on methods to control the mechanisms and reduce their impact on the drift of sensors in order to increase the reliable operational life of a base metal thermocouples. This led to a new thermocouple design that addresses the issue of depletion/change in composition due to the contamination from the sheath to the thermoelements at high temperatures. SOLVING THE PROBLEM With the information that had been gathered by researchers it was considered that to reduce the drift in base metal mineral insulated thermocouples it would be necessary to reduce the transfer of elements from the outer sheathing across the mineral insulation (MgO) to the thermocouple conductors. As the contaminating elements that were causing the impact on drift had been established, the simple answer would be to employ an outer sheathing that did not contain the contaminants or had a significantly reduced level of these alloying elements while still being able to work in the demanding and wide ranging thermal environments employed in today s diverse industrial conditions. Currently used outer sheathing materials, such as Inconel 600, 310, 316, 304 stainless steels, are able to work in demanding operating environments but do contaminate significantly the thermoelements during operation. A new sheath is needed to reduce the potential of contamination of the thermocouple wires. THE NEW DESIGN: THE DOUBLE WALL The innovative new design researchers from the University of Cambridge proposed to reduce the drift was to put a barrier, in the form of an additional inner sheath, between the current outer sheath and base metal thermocouple conductors: this design will be called herein double wall or dual wall configuration and the resulting sensor double wall or dual wall thermocouples (Figure 1). The mineral insulated thermocouple cable sheathing has had a positive effect on reducing the contributions of oxidation and other physical/chemical changes by protecting the thermocouple conductors from the external or operational environments. However, the same outer sheathing has now become, particularly at 2

3 Two sets of test sensors of each design were manufactured for both constructions in type K and type N thermocouples. The tests were conducted in a calibration horizontal tube furnace with a 500 mm immersion depth. The operating atmosphere was air. The test thermocouples were subjected to a program of thermal cycling over the temperature range 450 to 1250 C (842 to 2282 F). The time and temperature profile used for a cycle is shown in Figure 2 Each thermal cycle took just below 400 minutes to complete. Thermal calibration of the test sensors was conducted using standard type R calibrated reference thermocouples and a Datron Wavetek 1271 DMM as the measurement device. All thermocouple readings were referenced to 0 C (32 F) and all measurements were also referenced to the ITS 90 temperature scale. Figure 1 This led to work being required to find or develop a suitable sheathing that could be used in the manufacturing process of mineral insulated cable and could also survive the operational industrial thermal environments it would be used in, while stopping or reducing the transfer of elements from the outer sheath that creates the thermocouple conductor inhomogeneity and the resultant drift. After significant investigation and testing a suitable alloy that had the required properties was identified. Patents were filed and manufacture of prototype cables was undertaken by using conventional thermocouple manufacturing technology. LABORATORY RESULTS A detailed program of laboratory and industrial testing was then conducted to prove the concept. The following are the latest results from a sequence of testing undertaken in the CCPI Europe Ltd. UKAS accredited (Lab. No 0600) calibration laboratory. The tests were conducted on both type K and type N Inconel 600 sheathed thermocouples manufactured using the same core batch material for both the conventional mineral insulated cable construction and the new design dual wall low drift construction. Due to the dual wall cable being a prototype it was only available as a 3.2 mm overall diameter Inconel 600 sheathed cable. The conventional construction used for tests was a 3.0 mm diameter Inconel 600 sheathed cable. For the type K thermocouple tests calibrations of the test sensors were conducted on the 1 st cycle, 16 th, 31 st, 46 th 61 st, 76 th and 91 st cycles respectively. Identical tests were conducted on the type N test sensors on the 1 st cycle, 30 th, 60 th and 90 th cycles. These tests were conducted under laboratory conditions and the total combined test time was just under 25 days. The test thermocouples were not subjected to any mechanical stress during the tests. Uncertainties of measurement for the calibrations were estimated to be between 1 and 1.4 C (1.8 and 2.5 F) with a coverage factor k=2, providing a level of confidence of approximately 95%. The uncertainty evaluation has been carried out in accordance with International requirements. TYPE K SENSORS The results obtained for the type K test sensors are shown in Figure 3, 4 and 5. Figure 3 shows the results for the conventional type K mineral insulated construction. Between the 1 st and 16 th cycle an initial lowering or shift in calibration values of between +0.7 and -0.7 C (+1.26 to F) was experienced. After the 31 st cycle this drift had increased to values between -1.8 and -2.4 C (-3.2 to -4.3 F), by the 46 th cycle this had further increased to between -4.5 and -6.1 C (-8 and -11 F), by the 61 st cycle by between -7.1 and -10 C (-12.7 and F), by the 76 th cycle by between -9.1 and C (-16.4 and F) and finally after the 91 st cycle the overall change in calibration values was between-10.6 and -17 C (-19 and F). 3

4 Figure 4 shows the results for the dual wall type K mineral insulated construction. Between the 1 st and 16 th cycle an initial lowering or shift in calibration values of between +1.1 and +1.5 C (+1.9 to +2.7 F), slightly larger than that of the conventional construction, was experienced. However, after the 31 st cycle this drift had in fact slightly reduced to values between +0.7 and +1.4 C (+1.2 to +2.4 F), by the 46 th cycle this had further stabilised to be between +0.1 and +1 C (+0.2 and +1.7 F), by the 61 st cycle was by between +0.1 and +0.9 C (+0.1 and +1.5 F), by the 76 th cycle by between +0.4 and +2 C (+0.7 and +3.6 F) and finally after the 91 st cycle the overall change in calibration values was by between +0.5 and +2.9 C (+1 and +5.3 F). Figure 5 shows the total calibration shifts for both the conventional and double wall type K thermocouple cable designs between the 1 st and 91 st cycles. TYPE N SENSORS The results obtained for the type N test sensors are shown in Figure 6, 7, 8. Figure 6 shows the results for the conventional type N mineral insulated construction. Between the 1 st and 30 th cycle an initial lowering or shift in calibration values of between -1.1 and -2.2 C (-1.9 to -4 F) was experienced. After the 60 th cycle this drift had increased to values between -7.9 and C (-14.2 to F); by the 90 th cycle the conventional thermocouples were open circuit and had physically broken in the calibration furnace. Figure 7 shows the results for the double wall type N mineral insulated construction. Between the 1 st and 30 th cycle an initial lowering or shift in calibration values of between -0.4 and -0.7 C (-0.8 to -1.3 F) was experienced: this was smaller than that of the conventional construction. After the 60 th cycle this drift had changed to values between -2.1 and -2.7 C (-3.7 to -4.9 F) and by the 90 th cycle this had stabilised to be between 0 and -3.3 C (0 and -6 F). Figure 8 shows the total calibration shifts for both the conventional and double wall type N thermocouples between the 1 st and 60 th cycles. (Note that none of the tested conventional type N thermocouples were operational after 90 cycles). SUMMARY OF DUAL WALL CABLE DESIGN RESULTS The type K dual wall new design demonstrated an average 89.6% reduction in drift over all tests at all test temperatures after 91 cycles, with a maximum shift in calibration values between the 1 st and 91 st cycle being 2.9 C (5.2 F) against a maximum 17 C (30.6 F) value shown by the conventional mineral insulated cable design. The type N dual wall new design demonstrated an average 74.8% reduction in drift over all tests at all test temperature after 60 cycles (note that no comparison could be made at 90 cycles as the conventional mineral insulated cables were open circuit at 90 cycles), with a maximum shift in calibration values between the 1 st and 90 th cycle being 3.3 C (5.9 F). CONCLUSION The temperature cycle time and temperatures for this test were selected to give as close a simulation as possible to operational conditions when thermocouples are used for temperature uniformity surveys to meet aerospace heat treatment specifications such as AMS 2750 rev E, BAC 5621 rev K and RPS 953 issue 21. The test results showed that the conventional type K design maintained calibration limits and stayed within IEC : 2013 class 1 and ASTM E230 special tolerances up to the 16 th cycle. By the 31 st cycle the conventional design cable was outside both class 1 and special tolerances and by cycle 46 th cycle the results showed they had drifted outside both IEC : 2013 class 2 and ASTM E230 standard tolerances. Whereas the new design dual wall type K sensors maintained calibration values to meet both IEC : 2013 class 1 and ASTM E230 special tolerances for the full duration of the tests. The type N conventional mineral insulated cable construction performed better than the type K counterpart and maintained both IEC : 2013 class 1 and ASTM E230 special tolerances values up to and including the 30 th cycle. By 60 th cycle it had fallen outside class 2 and standard tolerances values. The new design dual wall type N construction sensors had, like the type K double wall version, maintained class 1 and special tolerance limits for the full test duration. The conclusion was that both the type K and type N conventional designs were showing calibration shifts or drift outside useable tolerances and therefore, after only a third of the test time had elapsed, were operating outside operational requirements required by the specifications. Both the type N and type K new dual wall designs showed significant reductions in calibration shift or drift and in fact maintained the required calibration tolerances throughout the full duration of the test cycles. INDEPENDENT REVIEW Interest in the new double wall cable design has been very high and its application for numerous high temperature, high duration, high accuracy uses in industry has been investigated. 4

5 Several European Metrology Institutes have shown interest in the double wall technology and in the near future the double wall thermocouples will be tested in independent test campaigns by the Metrology Institutes as part of joint research projects. THE NEXT STEP IN R&D The potential of the double wall thermocouple configuration has only been partially explored so far and further development is needed to capture entirely its benefits. An optimisation of the inner and outer wall thicknesses can lead to tailor to specific applications the performance of the double wall thermocouples in terms of both drift and life. The flexible design provided by the double wall thermocouples and the optimisation process also offer the opportunity to increase the thermoelements diameter initially used for the first prototypes of the double wall thermocouples: this potentially will have the effect of reducing drift rates further. STANDARDISATION The current data and results on double wall thermocouples will be brought to the attention of standardisation committees: standardisation will allow a large industrial community to profit from the improved performance provided by double wall thermocouples. Figure 2 5

6 Correction in Degrees C Correction in Degrees C Figure 3 +2 Type K 3.0mm Single wall calibration results on cycle 1, 16, 31, 46, 61, 76 and cycle /F2 3mm Single Wall cycle /B2 3mm Single Wall cycle /F2 3mm Single Wall cycle /B2 3mm Single Wall cycle /F2 3mm Single Wall cycle /B2 3mm Single Wall cycle /F2 3mm Single Wall Cycle /B2 3mm Single Wall Cycle /F2 3mm Single Wall Cycle /B2 3mm Single Wall Cycle /F2 3mm Single Wall Cycle /B2 3mm Single Wall Cycle /F2 3mm Single Wall Cycle /B2 3mm Single Wall Cycle 91 IEC class 1 tolerances Figure Type K 3.2mm Double wall calibration results on cycle 1, 16, 31, 46, 61, 76 and cycle /F2 3.2mm Double Wall cycle /B2 3.2mm Double Wall cycle /F2 3.2mm Double Wall cycle /B2 3.2mm Double Wall cycle /F2 3.2mm Double Wall cycle /B2 3.2mm Double Wall cycle /F2 3.2mm Double Wall Cycle /B2 3.2mm Double Wall Cycle /F2 3.2mm Double Wall Cycle /B2 3.2mm Double Wall Cycle /F2 3.2mm Double Wall Cycle /B2 3.2mm Double Wall Cycle /F2 3.2mm Double Wall Cycle /B2 3.2mm Double Wall Cycle 91 IEC Class 1 tolerances 6

7 Correction in Degrees C Differencen in Degrees C Figure Type K 3.2mm Double wall and 3.0mm Single wall change in Degrees C between cycle 1 and cycle /F2 3mm Single Wall 13222/B2 3mm Single Wall 13168/F2 3.2mm Double Wall 13168/B2 3.2mm Double Wall Figure Type N 3mm Single wall calibration results on cycle 1, 30 and cycle /F Single Wall 1st Cycle 13047/B Single Wall 1st Cycle 13047/F Single Wall 30th Cycle 13047/B Single Wall 30th Cycle 13047/F Single Wall 60 cycles 13047/B Single Wall 60 cycles IEC : Class 1 tolerances 7

8 Differencen in Degrees C Correction in Degrees C Figure Type N 3.2mm Double wall calibration results on cycle 1, 30, 60 and cycle /F Double Wall 1st Cycle 13167/B Double Wall 1st Cycle 13167/F Double Wall 30th Cycle 13167/B Double Wall 30th Cycle 13167/F Double Wall 60 cycles 13167/B Double Wall 60 cycles 13167/F Double Wall 90 cycles 13167/B Double Wall 90 cycles IEC Class 1 tolerances Figure 8 Type N 3.2 and 3.0mm Double and Single wall change in Degrees C between cycle 1 and cycle /F Single Wall 13047/B Single Wall 13167/F Double Wall 13167/B Double Wall -End- 8