|
Strain Surveys in a Spin Tester | |||||||||||||||||||||||||||||||||||||||||||||
![]() | |||||||||||||||||||||||||||||||||||||||||||||
|
Strain Surveys & Failure AnalysisIntroductionTest Devices provides customers with a Strain Survey Testing service for high speed rotational devices. Strain Surveys of static and linearly moving parts help engineers understand how their designs can be improved to increase life and performance, while reducing material costs. Strain Surveys for rotating parts are considerably more difficult due to the very high g load environment that both the DUT (device under test) and the instrumentation must withstand. The DUT is usually designed and subjected to FEA (Finite Element Analysis) which can locate and estimate nominal stress points. However, critical parts must be tested to verify actual stresses. Additionally, other points of concern can also be measured during Strain Surveys such as bolt holes, blades, attachment points, etc. which are very difficult to model.
Figure 1 TDIs robust equipment allows speeds all the way up to and including burst speeds. Unlike one-time brittle coatings sometimes used, Test Devices can incorporate cycling the DUT multiple times and vary the test temperatures resulting in a Strain Survey which more accurately represents design performance. Applications
Figure 2 Reasons for Testing
What Are the Advantages of a Strain Survey?1. The higher the operational speeds, the more important understanding the strain of a part becomes. The need for strain information (data) becomes more critical if the parts are to be spun-up and spun-down often rather than running in a steady state condition. Strain calculations are generally done in an FEA model to quickly optimize the model. However, a model is just that, a model, and real world performance varies somewhat from that baseline. Because of the energy involved in spinning parts, misjudgments can be dangerous and it is advantageous to have empirical knowledge in addition to the theoretical values. By their nature, odd shapes, welds, fillets, etc are difficult to analyze. These are the areas where stresses concentrate, and potentially lead to failure. Thus, a fuller understanding of how these locations affect the parts overall reliability is prudent. 2. FEA software programs are quite good at modeling one piece part designs. However, they do not do as well where parts can move relative to one another, such as assemblies. Here is where empirical testing can bring forth issues that do not show up in modeling. Should these issues not be addressed and find their way into the field, they can be quite expensive in both cost and reputation. 3. Structures incorporating composite are not as well understood as those which use metals and can not be adequately modeled with FEA due to inconsistent material properties. Here, a strain survey can tell the designer information about the part not attainable in any other manner. Typically, composite structures are made up of layers of fabric which can shift during rotation creating stress and strains that can not be predicted. A Strain Survey can provide data not only on the product but the quality of the manufacturing process, as well. 4. When parts are spun at high speed they grow as a result of the high centrifugal forces. In a perfect world, the material would recover its original dimensions when the speed is reduced. In fact, this does not happen and some of the growth is permanent. The ratio of recoverable growth to non-recoverable growth is part of an Elastic / Plastic analysis. Good data for this analysis is vital in determining the useful life of a part. 5. Besides knowing the rotational strain, there are also vibratory and thermal strains superimposed onto a part. Dynamic Strain Surveys can measure the damping inherent in a system to minimize vibration and provide information on the effectiveness of damping added to an assembly. Similarly, thermal gradient strain can be measured with an appropriate setup and gauges. 6. The clamping force of a bolt is proportional, but not necessarily directly related to the torque applied and is difficult to determine. Strain gauging bolts can calibrate the effects of friction, galling, lubrication, etc., and provide a more accurate conversion of clamping force from torque. 7. In the world of jet aircraft, failures at inopportune times can be very serious. Strain surveys can provide information on the propagation of cracks while rotating or cycling from start through operating speeds and back (LCF). This can provide valuable information on imminent failure risk . 8. Strain Surveys are very cost effective. Often, there are only a few areas of concern in a component. TDI can usually instrument a test piece with 6 strain gauges and run the test less expensively than a complete FEA analysis. Not only does this save money but the customer has data which is more dependable than that produced from a model, because the analysis has been taken one step further - real information on a real part.
Figure 3 In the graphic above, sensors positioned on the DUT were used to calibrate the FEA model shown just below the DUT. A profile of the DUT over the designers chosen range of speeds was created from the resulting data. In another example, Test Devices was able to extrapolate the burst speed to within 100 rpm from test data generated using high elongation sensors at a considerable savings over burst testing. Identifying the areas of concern in a component as early as possible in a design can save substantial dollars for a program. The US Military stated at a conference that the cost of an unidentified problem multiplies from the start to the end of a program by the following multipliers found in design = 1x, found in prototypes = 10x, found in certification = 100x, found in a field failure = 1000x. Types of TestsThere are three broad categories of Strain Surveys: Cold Tests, Ambient Tests and Heated Tests. Though there are variations, all Strain Surveys are essentially customized for the component to be tested. Each test must be instrumented precisely to be sure that the information provided by the test is accurate, germane and representative. Ensuring that the data is good requires consideration of several factors: the material properties of the component under test, the capabilities of the adhesives used to retain the sensors and the backing of strain gauge itself. Installing the wrong kind of gage for a specific application or improperly installing gages can give misleading and inaccurate data. All these properties must be matched for optimal results. Test Devices has used strain gauges embedded into fabric windings of composite materials. Inter-laminar strains can be measured in products, such as flywheels, alerting one to an overspeed condition that could destroy the part. A Strain Survey can be done in conjunction with other tests such as a burst test where the strain can be measured accurately up to the point of destruction. The TestA typical Strain Survey test involves spinning the instrumented disk (i.e. with strain gages mounted), a proper containment structure (safety in case of disk failure), and a slip ring (to get data to and from the spinning disk). To perform the test, Test Devices will do the following:
Figure 4 Design overview:
Installation:
DefinitionsReduced Data (aka data reduction) - Data collected from various sensors are written to a file in raw (native) format. For example, in millivolts or ohms. While accurate, these data are very difficult to interpret. The data reduction process will convert the raw data into the appropriate engineering units (i.e. millivolts to thousands of an inch using a calibrated curve, etc.) and present the data in a form (charts, histograms, etc.) that can be more easily analyzed and interpreted. Depending on the data to be collected this can be either real time or post-processed. (return) Gauges, gages & sensors describe the same thing and are used interchangeably in the industry.(return) LCF = Low Cycle Fatigue. Much like a paper clip is fatigued when it is bent back an forth in your hand, a high speed rotor fatigues as it cycles from low to high speeds until a point at which it fails. [see LCF page] (return) Composite used here refers to some type of filament woven into a fabric which in turn is encased in a binder such as epoxy. Carbon and glass fiber composites being the most common.(return) AWG = American Wire Gauge (Hardwarebook.net or pupman.com).(return) For additional information of answers to specific questions, please contact us.
|
|
|
Copyright© 2003, 2004, 2005, 2006, 2007 Test Devices Inc  |  571 Main Street  |  Hudson, MA 01749-3035 U.S.A. |