Table of Contents ECNDT '98
Session: Automotive Industry
Helium Leak Test at 950°CG. Schröder, F. Pauly, D. Grunwald,
Forschungszentrum Jülich GmbH, Central Department of Technology
D - 52425 Jülich, Germany
J.d.C. Payão Filho
|TABLE OF CONTENTS|
The leak detection methods used in industry and research range from the detection of a dripping valve up to a detectable gas loss of 1 cm³ over a period of 300 years! This corresponds to a difference in leak rate of about 12 orders of magnitude. Among the known leak detection methods, the leak test with helium as the tracer gas has become the most universal method due to its high sensitivity and versatile application. State-of-the-art equipment enables the reliable detection of leaks with leak rates in the range of 1 to 10-10 mbar·l·s-1, detection being normally performed at room temperature and a differential pressure of p = 1 bar.
In the course of developing a high-temperature solid oxide fuel cell it was necessary to examine components for leak tightness under their later operating conditions of 950 °C and only 200 mbar differential pressure.
These components included special connectors for later gas supply. The connectors consisted of two concentric metallic tubes. Due to differently selected thermal expansion coefficients, the expansion of the inner tube was greater than that of the encasing tube. This resulted in a sealing effect at high temperatures merely due to this thermally induced contact pressure, which had to be quantified. This was done varying both the gap width and the surface finish.
For the experiments, a special vacuum-tight test vessel heatable to 950 °C was built to accommodate one test specimen at a time. The vessel was evacuated during the experiments. The test specimens were exposed to helium flowing through thin pipes leading outwards. A differential pressure of exactly 200 mbar was adjustable by means of a special valve cycle and pre-evacuated intermediate vessels. This made it possible to determine the leak rate on red-hot components.
In the case of a dripping tap it can be seen with the naked eye: it is leaky and liquid penetrates through the leak! However, leaks of this magnitude are not permissible for almost any technical application. As the leaks become smaller, the amount of liquid penetrating through them also decreases. Leakage is stopped when the leak is clogged by large liquid molecules.
The leak rate is measured in mbar·l·s-1. A test object can be regarded as "fluid-tight" below a (helium) leak rate of 10-1 mbar.l·s-1. For gases with their smaller atoms, however, leaks of this size are still huge holes whose leak rates are not tolerable in most state-of-the-art applications. Large amounts of gas can still escape even through a leak whose size is only about one millionth of the fluid-tight leak. In the case of a leak rate of 10-8 mbar·l·s-1 and a pressure difference of p = 1 bar, for example, the gas loss still amounts to 0.3 cm³ per year. In many cases, however, even these low leak rates are still much too high for special applications.
Helium is used as the tracer gas for detecting these small leaks. Helium is nontoxic, does not react with other elements and can even penetrate through the smallest leaks due to its very small atoms. Moreover, helium is only contained in very small quantities (5 ppm) in the ambient air so that air does not substantially disturb the measurement.
Efficient helium detectors can detect leaks in a measuring range from about 1 to 10-10 mbar·l·s-1, i.e. over ten orders of magnitude. Moreover, a helium leak rate of 10 -10 mbar·l ·s-1, for example, only corresponds to a gas loss of about 1 cm³ in 300 years! However, about 250 million gas atoms per second still flow even through this unimaginably small leak!
There are different test arrangements for leak detection. The test specimen is either externally sprayed with helium and the interior evacuated by the leak detector or the test specimen is encased and the interior exposed to helium. In the latter case, the outer encasing is connected to a leak detector. In both cases, the helium escaping through a possible leak is recorded by a special measurement cell (helium mass spectrometer) in the leak detector. The method to be used largely depends on the object to be tested and on the expected information concerning tightness. If the leak is to be localized (= local testing), the test arrangement is different from that required for a statement on the overall tightness of a test specimen (= integral testing).
All test variants are normally operated at a pressure difference of p = 1 bar and at room temperature. In special cases, however, it may be necessary to specify the tightness of a component under special operating conditions. This will then require a special test arrangement.
The Research Centre Jülich is concerned with the development of fuel cells for future energy supply. In a fuel cell, the bound energy of stored hydrogen is directly converted into electric current with the aid of oxygen. Depending on the fuel cell design, this may involve very high operating temperatures. In the course of these developments, the tightness of various components had to be examined under their future operating conditions, i.e. at 950 °C and a differential pressure of only 200 mbar. The special test arrangement required for the leak test will be described in the following using an example.
The objects to be examined are tube connectors for gas supply in a high-temperature solid oxide fuel cell. These connectors of very simple design (Fig. 1) consist of two steel tubes which can be loosely fitted into one another in the cold state. The inner tube is made of 1.4841 (X 15 CrNiSi 25 20) material with a thermal expansion coefficient of = 19.0 · 10-6 m/m°C and the outer tube of 1.4742 (X CrAl 18) with = 13.5 · 10-6 m/m °C.
When heated, the inner tube expands more than the outer tube due to its greater thermal expansion coefficient. Both parts are thus pressed so tightly onto each other that a sealing effect is achieved. When the fuel cell is placed out of operation, this connection can be very easily disengaged after cooling.
|Fig 1: Object to be tested: connector|
The aim of the investigations was to quantify the leak rate of this pressure connection at high temperatures up to 950 °C. For this purpose, both the gap between the two tubes and the surface roughnesses of the tubes were varied.
In order to exclude corrosion influences and scaling on the connector surfaces, all tests had to be carried out in vacuum.
The experimental setup is schematically shown in Fig. 2.
|Fig 2: Schematic representation of the experimental setup|
|1||heater||7||two temperature measuring points|
|2a||vacuum vessel||8||helium leak detector|
|2b||lid||9||PC, data acquisition|
|3||specimen, object to be tested||10||helium storage tank|
|4||thermal insulation||11||mechanical forepump|
|5||cooling for O-ring seal||12||pressure (vacuum) measuring instrument|
|6||connecting pipe to the specimen||V||...valve|
The specimen to be examined (3) is placed on a frame in the vacuum vessel (2a), which stands upright in the heater (1). The heater has a connected electrical power of 5.6 kW. The vacuum vessel made of heat-resistant INCONEL can be closed vacuum-tight with a lid (2b). A thermal insulation (4) provided approximately at mid-height in the vacuum vessel reduces the temperature in the region of the lid. Additional water cooling (5) ensures that the permissible operating temperature of the O-ring seal in the lid is not exceeded. The vacuum vessel is connected with the helium leak detector (8). The vacuum is produced by the pumps of the leak detector.
The interior of the specimen (3) is connected with a small helium storage tank (10) by two thin (Ø 2 mm) pipes (6). The pipes (6) and the helium storage tank (10) are evacuated by a small vacuum pump (11) during the heating phase and between the individual measuring points, so that the same pressure of pabs = 0 bar prevails in the vacuum vessel and in the specimen. The interior of the specimen is only exposed to helium when the temperatures are reached at which the leak rate is to be determined. This is the case at 400, 500, 600, 700, 800 and 950 °C both during the heating phase and during cooling.
|Fig 3: Experimental setup|
In order to determine the leak rate, valves V1 and V3 are closed and the helium storage tank is filled with helium up to an absolute pressure of 200 mbar. Subsequently, valves V2 and V4 are closed and valve V1 is opened. This ensures a 100 % helium concentration in the specimen. In the absence of any major untightness, a differential pressure of p = 200 mbar is established between specimen and vacuum vessel during the measurement. The helium penetrating through the sealing gap on the specimen flows to the helium leak detector (8) where it is recorded as the integral leak rate.
Fig. 3 shows a photograph of the experimental setup. On the left you can see the heater with built-in vacuum vessel. The LEYBOLD UL 100-plus helium leak detector can be seen in the centre on the right.
Since this report concentrates on the special testing technique applied here, the result of one test series will only be presented as an example in the following. Fig. 4 shows the leak rate curves for four different specimens at different temperatures during one heating and one cooling cycle. As was to be expected, the leak rate decreases with rising temperature, i.e. with increasing pressure per unit area in the sealing gap. Helium leak rates in the range from 10-2 to 10-3 mbar·l·s-1 were measured at 900 to 950 °C. The different leak rates of the specimens are attributable to different gap widths and surface roughnesses.
|Fig 4: Experimental results of four specimens|
The measured leak rates are composed of the gas loss (leak rate) in the sealing gap of the specimen (3) and of the helium permeation through the material walls. Permeation takes place through the walls of the two connecting pipes (6) leading from the specimen through the lid (2b) of the vacuum vessel (2a) outwards as well as through the walls of the specimen (3).
A test was carried out to determine the leak rate through the connecting pipes (6) to the specimen, which will not exist in the actual fuel cell. For this purpose, the connecting pipes were built into the test vessel without a specimen, heated and then exposed to helium from the inside. However, no helium leak rate due to permeation significantly influencing the results was measured even at 950 °C for a prolonged test time of approx. two hours.
A special experimental setup has been conceived to investigate the leak tightness of tube connectors for gas supply at high-temperature solid oxide fuel cells. In this way, it was possible to carry out a helium leak test under future operating conditions, i.e. at 950 °C and a differential pressure of 200 mbar.
The measurements revealed leak rates in the range from 10-2 to 10-3 mbar·l·s-1 at 950 °C. The leak rate can be further optimized by varying the gap and surface finish. Helium permeation through material walls does not play any significant role. The tests have shown that the connectors are suitable for use at 950 °C.